It's nice to see a paper that confirms what anyone who has practiced using LLM tools already knows very well, heuristically. Keeping your context clean matters, "conversations" are only a construct of product interfaces, they hurt the quality of responses from the LLM itself, and once your context is "poisoned" it will not recover, you need to start fresh with a new chat.
Helmut10001 · 10h ago
My experiences somewhat confirm these observations, but I also had one that was different. Two weeks of debugging IPSEC issues with Gemini. Initially, I imported all the IPSEC documentation from OPNsense and pfSense into Gemini and informed it of the general context in which I was operating (in reference to 'keeping your context clean'). Then I added my initial settings for both sides (sensitive information redacted!). Afterwards, I entered a long feedback loop, posting logs and asking and answering questions.
At the end of the two weeks, I observed that: The LLM was much less likely to become distracted. Sometimes, I would dump whole forum threads or SO posts into it, when it said "this is not what we are seeing here, because of [earlier context or finding]. I eliminated all dead ends logically and informed it of this (yes, it can help with the reflection, but I had to make the decisions). In the end, I found the cause of my issues.
This somewhat confirms what some user here on HN said a few days ago. LLMs are good at compressing complex information into simple one, but not at expanding simple ideas into complex ones. As long as my input was larger than the output (either complexity or length), I was happy with the results.
I could have done this without the LLM. However, it was helpful in that it stored facts from the outset that I had either forgotten or been unable to retrieve quickly in new contexts. It also made it easier to identify time patterns in large log files, which helped me debug my site-to-site connection. I also optimized many other settings along the way, resolving not only the most problematic issue. This meant, in addition to fixing my problem, I learned quite a bit. The 'state' was only occasionally incorrect about my current parameter settings, but this was always easy to correct. This confirms what others already saw: If you know where you are going and treat it as a tool, it is helpful. However, don't try to offload decisions or let it direct you in the wrong direction.
Overall, 350k Tokens used (about 300k words). Here's a related blog post [1] with my overall path, but not directly corresponding to this specific issue. (please don't recommend wireguard; I am aware of it)
Recently, Gemini helped me fix a bug in a PPP driver (Zephyr OS) without prior knowledge of PPP or even driver development really. I would copy-paste logs of raw PPP frames in HEX and it would just decode everything and explain the meaning of each bytes. In about an hour, I knew enough about PPP to fix the bug and submit a patch.
I’m not saying that your approach is wrong. But most LLM workflows are either brute forcing the solution, or seeking a local minima to be stuck in. It’s like doing thousands of experiments of objects falling to figure out gravity while there’s a physics textbooks nearby.
Ironically, I could’ve read all 50 pages of that RFC and still missed the actual issue. What really helped was RFC 1331[0], specifically the "Async-Control-Character-Map" section.
That said, I’m building a product - not a PPP driver - so the quicker I can fix the problem and move on, the better.
I could also walk everywhere, but sometimes technology can help.
There’s no way I could fully read that RFC in an hour. And that’s before you even know what reading to focus your attention on, so you’re just being a worse LLM at that point.
tralarpa · 5h ago
Interesting that it works for you. I tried several times something similar with frames from a 5G network and it mixed fields from 4G and 5G in its answers (or even from non-cellular network protocols because they had similar features as the 5G protocol I was looking at). Occasionally, the explanation was completely invented or based on discussions of planned features for future versions.
I have really learned to mistrust and double check every single line those systems produce. Same for writing code. Everything they produce looks nice and reasonable on the surface but when you dig deaper it falls apart unless it's something very very basic.
Helmut10001 · 8h ago
Yes, it fells like setting the `-h` flag for logs (human readable).
Benjammer · 9h ago
That's some impressive prompt engineering skills to keep it on track for that long, nice work! I'll have to try out some longer-form chats with Gemini and see what I get.
I totally agree that LLMs are great at compressing information; I've set up the docs feature in Cursor to index several entire large documentation websites for major libraries and it's able to distill relevant information very quickly.
sixtyj · 4h ago
In Gemini, it is really good to have large window with 1M tokens. However, around 100,000 it starts to make mistakes and refactor its own code.
Sometimes it is good to start new chat or switch to Claude.
And it really helps to be very precise with wording of specification what you want to achieve. Or repeat it sometimes with some added request lines.
GIGO in reality :)
johnisgood · 1h ago
Oh my, I hate it when it rewrites >1k LOC. I have to instruct it to "modify only ..., do not touch the rest" and so forth, but GPT does not listen to this often, Claude does. I dunno about Gemini.
bentt · 11m ago
Yes even when coding and not conversing I often start new conversations where I take the current code and explain it new. This often gives better results than hammering on one conversation.
This feels like something that can be fixed with manual instructions which prompt the model to summarize and forget. This might even map appropriately to human psychology. Working Memory vs Narrative/Episodic Memory.
jimmySixDOF · 16m ago
>"conversations" are only a construct of product interfaces
This seems to be in flux now due to RL training on multiturn eval datasets so while the context window is evergreen every time, there will be some bias towards interpreting each prompt as part of a longer conversation. Mutliturn post training is not scaled out yet in public but I think it may be the way to keep on the 'double time spent on goal every 7 months curve'
morsecodist · 10h ago
This matches my experience exactly. "poisoned" is a great way to put it. I find once something has gone wrong all subsequent responses are bad. This is why I am iffy on ChatGPT's memory features. I don't notice it causing any huge problems but I don't love how it pollutes my context in ways I don't fully understand.
somenameforme · 7h ago
It's interesting how much the nature of LLMs fundamentally being self recursive next token predictors aligns with the Chinese Room experiment. [1] In such experiment it also makes perfect sense that a single wrong response would cascade into a series of subsequent ever more drifting errors. I think it all emphasizes the relevance of the otherwise unqualifiable concept of 'understanding.'
In many ways this issue could make the Chinese Room thought experiment even more compelling. Because it's a very practical and inescapable issue.
I don't think the Chinese room thought experiment is about this, or performance of LLMs in general. Searle explicitly argues that a program can't induce "understanding" even if it mimicked human understanding perfectly because programs don't have "causal powers" to generate "mental states".
This is mentioned in the Wikipedia page too: "Although its proponents originally presented the argument in reaction to statements of artificial intelligence (AI) researchers, it is not an argument against the goals of mainstream AI research because it does not show a limit in the amount of intelligent behavior a machine can display."
keiferski · 6h ago
Great comment on the Chinese room. That idea seems to be dismissed nowadays but the concept of “cascading failure to understand context” is absolutely relevant to LLMs. I often find myself needing to explain basic details over and over again to an LLM; when with a person it would be a five second, “no, I mean like this way, not that way” explanation.
OtherShrezzing · 3h ago
I find using tools like LMStudio, which lets you edit your chat history on the fly, really helps deal with this problem. The models you can host locally are much weaker, but they perform a little better than the really big models once you need to factor in these poisoning problems.
A nice middle-ground I'm finding is to ask Claude an initial conversation starter in its "thinking" mode, and then copy/paste that conversation into LMStudio and have a weaker model like Gemma pick-up from where Claude left off.
AstroBen · 8h ago
good point on the memory feature. Wow that sounds terrible
distances · 5h ago
The memory is easy to turn off. It sounded like a very bad idea to cross-contaminate chats so I disabled it as soon as ChatGPT introduced it.
b800h · 6h ago
I've been saying for ages that I want to be able to fork conversations so I can experiment with the direction an exchange takes without irrevocably poisoning a promising well. I can't do this with ChatGPT, is anyone aware of a provider that offers this as a feature?
stuffoverflow · 6h ago
Google AI studio, ChatGPT and Claude all support this. Google AI studio is the only one that let's you branch to a separate chat though. For ChatGPT and claude you just edit the message you want to branch from.
Garlef · 6h ago
Support: Yes.
But the UX is not optimized for this.
Imagine trying to find a specific output/input that was good in the conversation tree.
layer8 · 5h ago
Yes, it would be nice if you could at least bookmark a particular branch.
m4houk · 5h ago
I once built something like this for fun as a side project.
You can highlight some text in a chat and fork the chat to talk about that text selection, so the LLM has context of that along with the previous chat history and it responds in a new chat (entire chat history up to that point from the parent chat gets copied over - basically inspired by the Unix `fork`).
Your text selection from the parent chat would get turned into a hyperlink to the new child chat so you can always get to it again if you're reading the parent chat.
therockhead · 4h ago
I need to think about this a bit more, but I think I would love a thread feature in ChatGPT, so that it has the context up to the point of creation but doesn’t affect the main conversation. It would help in two ways, it keeps the main topic from getting poisoned , and allow me to minimise text clutter when i go off on tangents during the conversation.
lewdwig · 5h ago
T3.chat supports convo forking and in my experience works really well.
The fundamental issue is that LLMs do not currently have real long term memory, and until they do, this is about the best we can do.
bambax · 5h ago
On Openrouter you can delete previous answers (and questions) and maintain a separate conversation with different models.
But it would indeed be nice to either disable answers (without deleting them) or forking a conversation. It wouldn't be hard to implement; I wonder if there's a market for just this?
granra · 6h ago
Some 3rd party UIs offer this, I use typingmind sometimes that does but AFAIK some open source ones do too.
a_e_k · 6h ago
If you're happy running local models, llama.cpp's built-in web-server's interface can do this.
anonexpat · 6h ago
I believe Claude has forking in their web interface.
actualwitch · 4h ago
I stumbled upon this issue myself when designing prompts for agentic systems and got mad at the lack of tools to support this flow, so I built one myself! I called it Experiment, it allows easy conversation forking and editing while retaining all logs.
An interesting little example of this problem is initial prompting, which is effectively just a permanent, hidden context that can't be cleared. On Twitter right now, the "Grok" bot has recently begun frequently mentioning "White Genocide," which is, y'know, odd. This is almost certainly because someone recently adjusted its prompt to tell it what its views on white genocide are meant to be, which for a perfect chatbot wouldn't matter when you ask it about other topics, but it DOES matter. It's part of the context. It's gonna talk about that now.
dragonwriter · 6h ago
> This is almost certainly because someone recently adjusted its prompt to tell it what its views on white genocide are meant to be
Well, someone did something to it; whether it was training, feature boosting the way Golden Gate Claude [0] was done, adjusting the system prompt, or assuring that it's internet search for contextual information would always return material about that, or some combination of those, is neither obvious nor, if someone had a conjecture as to which one or combination it was, easily falsifiable/verifiable.
Source [0]. The examples look pretty clearly like they stuck it in the context window, not trained it in. It consistently seems to structure the replies as though the user they're replying to is the one who brought up white genocide in South Africa, and it responds the way that LLMs often respond to such topics: saying that it's controversial and giving both perspectives. That's not behavior I would expect if they had done the Golden Gate Claude method, which inserted the Golden Gate Bridge a bit more fluidly into the conversation rather than seeming to address a phantom sentence that the user supposedly said.
Also, let's be honest, in a Musk company they're going to have taking the shortest possible route to accomplishing what he wanted them to.
> This is almost certainly because someone recently adjusted its prompt to tell it what its views on white genocide are
Do you have any source on this? System prompts get leaked/extracted all the time so imagine someone would notice this
Edit: just realized you’re talking about the Grok bot, not Grok the LLM available on X or grok.com. With the bot it’s probably harder to extract its exact instructions since it only replies via tweets. For reference here’s the current Grok the LLM system prompt: https://github.com/asgeirtj/system_prompts_leaks/blob/main/g...
9dev · 7h ago
Well, telling an AI chatbot to insist on discussing a white genocide seems like a perfectly Elon thing to do!
lenkite · 4h ago
Probably because it is now learning from a lot of videos posted on X by misc right-wingers showing rallying cries of South African politicians like Julius Malema, Paul Mashatile etc. Not very odd.
nah, llms don't learn like this — they specifically added it to the system prompt
stevedonovan · 6h ago
Ah, Elon paying attention to hid companies again!
Context poisoning is not a uniquely LLM problem
ezst · 7h ago
The heck??
Adambuilds · 4h ago
I agree—once the context is "poisoned," it’s tough to recover. A potential improvement could be having the LLM periodically clean or reset certain parts of the context without starting from scratch. However, the challenge would be determining which parts of the context need resetting without losing essential information. Smarter context management could help maintain coherence in longer conversations, but it’s a tricky balance to strike.Perhaps using another agent to do the job?
unshavedyak · 10h ago
Has any interface implemented a .. history cleaning mechanism? Ie with every chat message focus on cleaning up dead ends in the conversation or irrelevant details. Like summation but organic for the topic at hand?
Most history would remain, it wouldn’t try to summarize exactly, just prune and organize the history relative to the conversation path?
ithkuil · 6h ago
"Every problem in computer science can be solved with another level of indirection."
One could argue that the attention mechanism in transformers is already designed to do that.
But you need to train it more specifically with that in mind if you want it to be better at damping attention to parts that are deemed irrelevant by the subsequent evolution of the conversation.
And that requires the black art of ML training.
While thinking of doing this as a hack on top of the chat product feels more like engineering and we're more familiar with that as a field.
nosefurhairdo · 10h ago
I've had success having a conversation about requirements, asking the model to summarize the requirements as a spec to feed into a model for implementation, then pass that spec into a fresh context. Haven't seen any UI to do this automatically but fairly trivial/natural to perform with existing tools.
hobofan · 6h ago
Not a history cleaning mechanism, but related to that, Cursor in the most recent release introduced a feature to duplicate your chat (so you can saveguard yourself against poisoning and go back to and unpoisoned point in history), which seems like an addmision of the same problem.
olalonde · 9h ago
Not sure if that's what you mean but Claude Code has a /compact command which gets triggered automatically when you exceed the context window.
I mean, you could build this, but it would just be a feature on top of a product abstraction of a "conversation".
Each time you press enter, you are spinning up a new instance of the LLM and passing in the entire previous chat text plus your new message, and asking it to predict the next tokens. It does this iteratively until the model produces a <stop> token, and then it returns the text to you and the PRODUCT parses it back into separate chat messages and displays it in your UI.
What you are asking the PRODUCT to now do is to edit your and its chat messages in the history of the chat, and then send that as the new history with your latest message. This is the only way to clean the context because the context is nothing more than your messages and its previous responses, plus anything that tools have pulled in. I think it would be sort of a weird feature to add to a chat bot to have the chat bot, each time you send a new message, go back through the entire history of your chat and just start editing the messages to prune out details. You would scroll up and see a different conversation, it would be confusing.
IMO, this is just part of prompt engineering skills to keep your context clean or know how to "clean" it by branching/summarizing conversations.
rrr_oh_man · 8h ago
Or delete / edit messages in AI Studio or Open Router.
QuadmasterXLII · 8h ago
the problem is that it needs to read the log to prune the log, and so if there is garbage in the log, which needs to be pruned to keep from poisoning the main chat, then the garbage will poison the pruning model, and it will do a bad job pruning.
kqr · 9h ago
Isn't this what Claude workbench in the Anthropic console does? It lets the user edit both sides of the conversation history.
freehorse · 2h ago
Which is why I really like zed's chat UX experience: being able to edit the full prior conversation like a text file, I can go back and clean it up, do small adjustments, delete turns etc and then continue the discussion with a cleaner and more relevant context.
I have made zed one of my main llm chat interfaces even for non-programming tasks, because being able to do that is great.
CompoundEyes · 9h ago
Agreed poisoned is a good term. I’d like to see “version control” for conversations via the API and UI that lets you rollback to a previous place or clone from that spot into a new conversation. Even a typo or having to clarify a previous message skews the probabilities of future responses due to the accident.
mh- · 9h ago
"Forking" or "branching" (probably better received outside of SWEs) a conversation really ought to be a first class feature of ChatGPT et Al.
HaZeust · 9h ago
It is in Google Gemini, which I really hate to say - but I've been using a lot more than GPT. I reckon I'll be cancelling my Pro if Gemini stays with this lead for my everyday workflows.
AI Studio is borderline unusable for long conversations. I don't know what in the world it's doing but it sure looks like a catastrophic memory leak in the basic design.
voidspark · 8h ago
I have been using it up to 100k tokens so far without issues. Never needed to go further than that. But much of that was in uploaded documents.
drittich · 8h ago
Also exists in LM Studio.
wunderwuzzi23 · 6h ago
This was part of ChatGPT from pretty much the beginning, maybe not the initial version but few weeks later- don't recall exactly
layer8 · 7h ago
This has been in ChatGPT from pretty early on? Just edit any prompt, it creates a new branch, and you can switch back and forth.
b800h · 6h ago
Blimey, I didn't realise the entire thread was saved when you edited a prompt. Very good! Mind you, it feels "unsafe". I'd like to be able to clone a thread.
gdudeman · 7h ago
It is!
It exists in Claude as a true branch - you can see the old threads - and in ChatGPT as without the history.
Edit a previous reply and hit “go” to see it in action.
gdudeman · 7h ago
This exists in Claude. Edit any previous message and it will fork the conversation.
amelius · 6h ago
I suppose that the chain-of-thought style of prompting that is used by AI chat applications internally also breaks down because of this phenomenon.
djmips · 8h ago
Happens with people too if you think about it.
kfarr · 6h ago
Who gets lost in multi-turn conversations?
TheOtherHobbes · 5h ago
Everyone?
How often in meetings does everyone maintain a running context of the entire conversation, instead of responding to the last thing that was said with a comment that has an outstanding chance of being forgotten as soon as the next person starts speaking?
MattGaiser · 10h ago
Yep. I regretted leaving on memory as it is poisoned my conversations with irrelevant junk.
neom · 10h ago
You can go in and delete memory items
veunes · 6h ago
What surprised me is how early the models start locking into wrong assumptions
tmountain · 1h ago
I often ask the LLM for a concise summary of the discussion so far—formatted as a prompt. I then edit it appropriately and use it to start a new conversation without the baggage. I have found this to be a very effective technique, but I imagine it will be automated sometime soon.
No comments yet
Sharlin · 10h ago
Seems like this is an aspect of their well-known overconfidence and the inability to self-reflect and recognize they have to ask for more details because their priors are too low. If you look at the output of reasoning models, it’s clear that the idea of asking for clarification very rarely occurs to them – when they’re confused, it’s just endless speculation of what the user might have meant.
This, of course, has certain implications as to the wisdom of the idea of “replacing human programmers”, given that one of the hard parts of the trade is trying to turn vague and often confused ideas into precise specifications by interacting with the shareholders.
Terr_ · 10h ago
> inability to self-reflect
IMO the One Weird Trick for LLMs is recognizing that there's no real entity, and that users are being tricked into a suspended-disbelief story.
In most cases cases you're contributing text-lines for a User-character in a movie-script document, and the LLM algorithm is periodically triggered to autocomplete incomplete lines for a Chatbot character.
You can have an interview with a vampire DraculaBot, but that character can only "self-reflect" in the same shallow/fictional way that it can "thirst for blood" or "turn into a cloud of bats."
layer8 · 7h ago
Not to mention that vampires don’t reflect. ;)
Sharlin · 9h ago
This is a tired semantic argument that does not bring any insight into the discussion. A token-predictor could still be trained to predict the tokens “I’m not sure what you mean because of points x, y, and z; could you elaborate?”
dkdbejwi383 · 1h ago
How would an LLM “know” when it isn’t sure? Their baseline for truth is competent text, they don’t have a baseline for truth based on observed reality. That’s why they can be “tricked” into things like “Mr Bean is the president of the USA”
JustFinishedBSG · 35m ago
It would "know" the same way it "knows" anything else: The probability of the sequence "I don't know" would be higher than the probability of any other sequence.
ben_w · 35m ago
The answer is the same as how the messy bag of chemistry that is the human brain "knows" when it isn't sure:
Badly, and with great difficulty, so while it can just about be done, even then only kinda.
root_axis · 8h ago
It could be trained to say that, but it's not exactly clear how you would reinforce the absence of certain training data in order to emit that response accurately, rather than just based on embedding proximity.
jsnider3 · 1h ago
Seems easy. Have a set of vague requests and train it to ask for clarification instead of guessing.
timdiggerm · 1h ago
How does it identify what's vague?
simianwords · 6h ago
Why does it seem so hard to make training data for this? You can cook up a few thousands of training data and do an RLHF.
Terr_ · 9h ago
It means if you want something resembling a self-introspective theory of mind, you need to arrange the overall document to cohere to documents where such things are/appear-to-be happening.
This leads us to new questions: How can we characterize and identify real-world documents which fit? How can we determine what features may be significant, and which of those can be easily transplanted to our use-case?
simianwords · 6h ago
There are a lot of words but it feels like you have never really used LLM's (apologies for the bluntness).
We see LLM's introspecting all the time[1].
>Notably, DeepSeek-AI et al. report that the average response length and downstreamperformance of DeepSeek-R1-Zero increases as training progresses. They further report an “aha
moment” during training, which refers to the “emergence” of the model’s ability to reconsider
its previously generated content. As we show in Section 3.2, this reconsideration behaviour is
often indicated by the generation of phrases such as ‘wait, ...’ or ‘alternatively, ...’
Unless they show you the Markov chain weights (and I've never seen one that does), that's confabulation, not introspection.
sitkack · 8h ago
You are just doubling down on protecting your argument.
I operate LLMs in many conversational modes where it does ask clarifying questions, probing questions, baseline determining questions.
It takes at most one sentence in the prompt to get them to act this way.
bigcat12345678 · 8h ago
> It takes at most one sentence in the prompt to get them to act this way.
What is this one sentence you are using?
I am struggling to elicite clarification behavior form llms
mdemare · 42m ago
"Any questions before you start coding?"
sandspar · 8h ago
Could you share your prompt to get it to ask clarifying questions? I'm wondering if it would work in custom instructions.
roywiggins · 7h ago
Anthropic found that it Claude will pretend that it used the "standard" way to do addition- add the digits, carry the 1, etc- but the pattern of activations showed it using a completely different algorithm. So these things can role play as introspecting- they come up with plausible post-hoc explanations for their output- but they are still just pretending, so they will get it wrong.
So you can teach a model to sometimes ask for clarification, but will it actually have insight into when it really needs it, or will it just interject for clarification more or less at random? These models have really awful insight into their own capabilities, ChatGPT eg insists to me that it can read braille, and then cheerfully generates a pure hallucination.
roenxi · 6h ago
> Anthropic found that it Claude will pretend that it used the "standard" way to do addition- add the digits, carry the 1, etc- but the pattern of activations showed it using a completely different algorithm.
That doesn't mean much; humans sometimes do the same thing. I recall a fun story about a mathematician with synesthesia multiplying numbers by mixing the colours together. With a bit of training such a person could also pretend to be executing a normal algorithm for the purposes of passing tests.
frabcus · 6h ago
Even then the human doesn't know how they execute the algorithm, or mix the colours together - our conscious self-reflective mind has limits as to how far into our neural network weights it can reach. Can get further with lots of meditation, but it is still definitionally limited (in information theory terms).
jcims · 7h ago
I agree that it's a tired argument, but there appears to be two separate things being discussed in this little corner of HN. Clarity in the problem it's being asked to solve, and confidence that the answer it has is correct.
I can trivially get any of the foundational models to ask me clarifying questions. I've never had one respond with 'I don't know'.
chipsrafferty · 6h ago
I've gotten lots of responses like "with the information you provided, I cannot answer that. Can you provide more information?"
Which IMO is the name as "idk"
littlestymaar · 6h ago
It's not a tired argument, and not just a semantic one it's a foundational characteristic of LLM.
> A token-predictor could still be trained to predict the tokens “I’m not sure what you mean because of points x, y, and z; could you elaborate?”
This is entirely true, and the key insight is even right in your sentence but you don't seem to grasp it. “could still be trained”: you can train an LLM into doing whatever you want it to, but you have to train it specifically for that!
In the beginning of LLM we witnessed this impressive phenomenon where the LLM exhibited emergent capabilities (I'm particularly thinking about LLMs being few shots learners about stuff that wasn't in their training corpus). And these emergent capabilities legitimately raised the question about “how intelligent these things are, really”.
But for the past three years, the key lesson is that this kind of emergent effect is too small to be useful, and the focus has been put towards creating purposely built datasets (with tons of “artificial data”) to train the model to explicitly do things we want it to do. And it works pretty well, as models' capabilities kept improving at a fast pace (and in particular, I don't see would we couldn't overcome the problem highlighted by this paper, with more synthetic data specifically designed for multi-turn conversation). But their progress is now strictly limited by their makers' own intelligence. You cannot just scrap the web throw compute at the problem and expect emergent intelligence to occur anymore. It's more “simulated intelligence” than “artificial intelligence”, really.
og_kalu · 3h ago
It's definitely a tired and semantical one because as he said, it brings no insight and is not even good at the analogy level. I can't have a conversation with Dracula and Dracula can't make decisions that affect the real world, so LLMs already break key aspects and assumptions of the 'Document Simulator'.
Pre-trained LLMs will ask clarifying questions just fine. So I think this is just another consequence of post-training recipes.
bytepoet · 7h ago
The inability of LLMs of ask for clarification was exactly the flaw we encountered when testing them on open-ended problems, stated somewhat ambiguously. This was in the context of paradoxical situations, tested on DeepSeek-R1 and Claude-3.7-Sonnet. Blog post about our experiments: https://pankajpansari.github.io/posts/paradoxes/
btbuildem · 33m ago
> This, of course, has certain implications as to the wisdom of the idea of “replacing human programmers”
Ironically, working with a junior dev is a lot like this -- setting them on a task, then coming back later with dogs and flashlights to retrieve them from the deep woods they've inevitably lost themselves in by just forging ahead, making assumptions, and asking no questions.
voidspark · 8h ago
> inability to self-reflect and recognize they have to ask for more details because their priors are too low.
Gemini 2.5 Pro and ChatGPT-o3 have often asked me to provide additional details before doing a requested task. Gemini sometimes comes up with multiple options and requests my input before doing the task.
rrr_oh_man · 8h ago
That's a recent development for (imho) higher engagement and reduced compute.
voidspark · 8h ago
It's for higher quality of output. Better solutions. These are the state of the art reasoning models (subscription only, no free access) which are smarter.
It also mainly happens when the context is clear that we are collaborating on work that will require multiple iterations of review and feedback, like drafting chapters of a handbook.
I have seen ChatGPT ask questions immediately upfront when it relates to medical issues.
bandrami · 4h ago
Close. Higher engagement means the user is more invested and values the solution more.
The users are being engineered more than the models are, and this isn't the only example.
voidspark · 3h ago
Are you employed at Google or OpenAI? Are you working on these frontier models?
In the case of medical questions it needs to know further details to provide a relevant diagnosis. That is how it was trained.
In other cases you can observe its reasoning process to see why it would decide to request further details.
I have never seen an LLM just ask questions for the sake of asking. It is always relevant in the context. I don't use them casually. Just wrote a couple of handbooks (~100 pages in a few days). Generating tens of thousands of tokens per session with Gemini.
rrr_oh_man · 2h ago
typical patterns to look out for:
- "Should I now give you the complete [result], fulfilling [all your demands]?"
- "Just say [go] and I will do it"
- "Do you want either [A, B, or C]"
- "In [5-15] minutes I will give you the complete result"
...
voidspark · 2h ago
> "Do you want either [A, B, or C]"
That's an example of what I'm talking about. Watch the reasoning process produce multiple options. That's what it is trained to do. That is problem solving, not "engagement". It requires more compute, not less. You see that more with the expensive models.
> "In [5-15] minutes I will give you the complete result"
I haven't seen that before and I don't see how it's relevant.
rrr_oh_man · 1h ago
> That's an example of what I'm talking about. Watch the reasoning process produce multiple options. That's what it is trained to do. That is problem solving, not "engagement". It requires more compute, not less. You see that more with the expensive models.
Fair point. Thanks for standing your ground and arguing so matter-of-factly with me! Appreciate it.
bobsyourbuncle · 8h ago
Isn’t this relatively trivial to correct? Just like chain of thought reasoning replaces end tokens with “hmm” to continue the thought can’t users just replace the llm tokens whenever it starts saying “maybe they are referring to” with something like. “Let me ask a clarifying question before I proceed.”
Sharlin · 8h ago
Indeed, I was just about to edit my comment because the same occurred to me. Someone is probably going to try just that soon enough.
veunes · 6h ago
Real programmers spend a ton of time just figuring out what people actually want. LLMs still treat guessing as a feature
amelius · 5h ago
This cartoon needs an update for what an LLM came up with:
> and the inability to self-reflect and recognize they have to ask for more details
They're great at both tasks, you just have to ask them to do it.
roywiggins · 7h ago
You can certainly convince them to ask for details, but I'm not sure whether that makes them any good at knowing when exactly to ask vs just asking some percentage of the time regardless.
That is, does it actually know when it doesn't know, or are you just making it less confident overall, so it asks questions with no actual insight? Convincing a model to roleplay as someone who doesn't know things vs teaching a model to have insight into when it does and doesn't need clarification seems like a tough one.
aleksituk · 1h ago
This is very interesting and I like the conversation about not only the technology itself, but also about the importance of thinking about the interface as a user experience and where / how it fits the paradigm.
We've been working on a lot of data processing and generation tasks. We've been doing this using an API primarily, but sometimes I end up testing creating data in a chat window and I first chat through what the requirements are for the data analysis / processing and then once I'm done I would like the whole conversation to be then summarised into basically a one-prompt process so that I can re-use it (because I can't really process new inputs via the chat).
Even when you do manage to get it down to a single prompt you can use in a chat and then ask the chat to just keep producing new data (like imagine a blog post in certain style if the base content is given as input and I'm making like 20 of them). If you produce these in the chat, there's notable benefits in that if something is wrong with the blog post the chat suggests, you can immediately edit it. The trouble is that the context window starts becoming so big that the chat starts to forget what the original instruction is and eventually you do have to just create a new chat.
One way to solve for this is having a chat with selective memory where you keep a task in memory, but you have the chat forget/not-include all the generated data in the context so that it stays clean, but only bring it to the context if the user refers to it.
Has anyone else done data processing types of tasks in chats and had issues like this? Are there some other tools to use or tricks to do in chats?
t-kalinowski · 1h ago
This was the main reason I wrote promptdown. I want to be able to edit the full chat history every turn, and the append-only standard chat interfaces don't make that easy.
I tested this another 300 times with gpt-4.1 to remove those obtrusive "em-dashes" everyone hates. Tested a single-pass baseline vs TSCE, same exact instructions and prompt "Remove the em-dashes from my linkedin post. . .".
Out of the 300 tests, baseline failed to remove the em-dashes 149/300 times. TSCE failed to remove the em-dashes 18/300 times.
It works, all the data as well as the entire script used for testing is in the repo.
arnaudsm · 2m ago
That's a lot of kilo-watt-hours wasted for a find and replace operation.
Have you heard of text.replace("—", "-") ?
thegeomaster · 2h ago
I slightly tweaked your baseline em dash example and got 100% success rate with GPT-4.1 without any additional calls, token spend, or technobabble.
System prompt: "Remove every em-dash (—) from the following text while leaving other characters unchanged.\n\nReturn only the cleaned text."
User prompt: <prompt from tsce_chat.py filled with em dashes>
Temperature: 0.0
podgorniy · 5h ago
There is a noticable issue when one builds LLMs interfaces around single turn conversations. Majority people expect linear conversations.
I've built telegram bot http://t.me/experai_bot as univresal UI to LLMs (with somewhat reduced functionality) exactly around idea "non-reply message means new conversation". Wanna keep context? Keep replying to replies of bot. Non-power user strugge with this idea.
--
Also I observed that OpenAI models performed worse replying to the same questions (for example list of options in reply got shorter) even with smallest system message. That was the case with 3.5, 4o. Don't know how modern ones behave. That made me decide not to include any system messages by default Still I give option to add ones if you need. You can even toggle them to mix-and-match.
jumploops · 7h ago
It's amazing that branching/forking isn't a core aspect of the main chat tools.
You can edit responses, sure, but then a bunch of other context is lost.
My flow is basically:
1. plan
2. build
3. branch (into some feature/esoteric dependency issue)
4. goto #2
Prompt pruning/branching should be a first-class tool for any LLM usage.
jampekka · 4h ago
Google AI studio at least has this. I found at least that implementation quite confusing though, which may be a reason it's not implemented in more "consumer oriented" tools.
Capricorn2481 · 6h ago
I've been kicking around making this for a while. BetterChatGPT at least has some good ergonomics around deleting history. But I agree that branching is the next step.
zacksiri · 10h ago
I've been working on solving this with quite a bit of success, I'll be sharing more on this soon. It involves having 2 systems 1st system is the LLM itself and another system which acts like a 'curator' of thoughts you could say.
It dynamically swaps in / out portions of the context. This system is also not based on explicit definitions it relies on LLMs 'filling the gaps'. The system helps the llm break down problems into small tasks which then eventually aggregate into the full task.
simianwords · 6h ago
This is a great idea. What you are doing is a RAG over the chat.
In the future such a distinction in memory hierarchies will be more clear
- Primary memory in the training data
- Secondary memory in context
- Tertiary memory in RAG
cadamsdotcom · 9h ago
Sounds like an exciting idea.
May I suggest - put what you have out there in the world, even if it’s barely more than a couple of prompts. If people see it and improve on it, and it’s a good idea, it’ll get picked up & worked on by others - might even take on a life of its own!
You can see it's going from introduction, asking me for my name, and then able to answer question about some topic. There is also another example in the thread you can see.
Behind the scenes, the system prompt is being modified dynamically based on the user's request.
All the information about movies is also being loaded into context dynamically. I'm also working on some technique to unload stuff from context when the subject matter of a given thread has changed dramatically. Imagine having a long thread of conversation with your friend, and along the way you 'context switch' multiple times as time progresses, you probably don't even remember what you said to your friend 4 years ago.
There is a concept of 'main thread' and 'sub threads' involved as well that I'm exploring.
I will be releasing the code base in the coming months. I need to take this demo further than just a few prompt replies.
adrianm · 8h ago
This is a class of mental critic from the Emotion Machine.
adiadd · 9h ago
would be great to get more info on what you’re building - seems interesting!
zacksiri · 9h ago
I publish my findings on my youtube channel and my blog, you're welcome to have a look. Both links are in my profile.
layer8 · 7h ago
So, Map-Reduce-of-Thought?
zacksiri · 6h ago
You could say that! hahaha! I'm happy to see someone understand what it is.
permo-w · 10h ago
I feel like at this point the LLM space is just filled with people solving and resolving the same problems over and over
kristianp · 6h ago
Just like the llms in multi-turn conversations.
meroes · 8h ago
It’s herding cats, not “learning”, which is a fine situation for some parts of workflows.
dankwizard · 9h ago
And everyone loves to chime in with their own excellence in prompt engineering
giordanol · 39m ago
Would love to see metrics that isolate recovery behaviour (if any)
badmonster · 5h ago
Why do LLMs struggle so much with recovering from early wrong turns in multi-turn conversations — even when all prior context is available and tokenized?
Is it due to the model's training distribution (mostly single-shot completions), the way context windows are encoded, or an architectural bottleneck?
Feels like there's no dynamic internal state that evolves over the conversation — only a repeated re-parsing of static history. Has anyone seen work on integrating memory/state mechanisms that allow belief revision within a session, not just regurgitation of past tokens?
JohnKemeny · 5h ago
We shouldn’t anthropomorphize LLMs—they don’t “struggle.” A better framing is: why is the most likely next token, given the prior context, one that reinforces the earlier wrong turn?
bandrami · 4h ago
Because Markov chains propagate forward in time
vjerancrnjak · 3h ago
Imagine optimizing/training on a happy path.
When you generate future tokens, you're looking at history tokens that are happy.
So how can a model, given sad tokens, generate future happy tokens if it did not learn to do so?
The work you're looking for is already here, it's "thinking". I assume they include sad tokens in the dataset, produce "thinking", which should result in happy tokens coming after thinking tokens. If thinking is bad (by looking at following happy tokens), then it's punished, if good, then descent.
ranyume · 10h ago
I'd like more research done on context understanding other than NIAH. I don't believe LLMs support the context length companies say they support. But I need to know this to effectively use the tools. At least for coding.
Stuff like this:
1. Do: Best practice for X model is to include at max 10k lines of code + task + CONVENTIONS.md + architecture guidance. Only queue tasks for components that are fairly decoupled from the rest of the codebase (e.g. small modules).
2. Don't: Start a project without a clearly defined architecture in this format. Don't ask for tasks that require X amount of reading hops to understand the logic.
I find it frustrating that companies release their benchmaxxing without helping developers actually use their models. It's more ironic that some people think of these AIs as employees. Employees can work with their boss about the best way to achieve things! With LLMs you don't even know how to communicate with them and as a result their output is unreliable.
skydhash · 8h ago
You could swap those recommendations for programming without LLMs. Open any software engineering books and you’ll see a lot of good recommendations for building software.
overflow897 · 2h ago
I believe we're already using llms to evaluate llm output for training, I wonder if there's some variation of that which could be used to identify when one llm gets "stuck".
I guess chain of thought in theory should do that but having variations on prompt and context might behave differently?
jsemrau · 7h ago
That's no surprise. When I was working on game theory and agent reasoning I reached the same conclusion a year ago.
My conclusion was that context needs to be managed well for the LLMs to manage accuracy in replies. Also, it helps to have a planning process ("graph reasoning") before task execution because it guardrails the models thought process.
This also introduces a discussion on general use vs workflow agent implementations as in the former it is much more difficult to generalize all components in structuring effective ReAct patterns.
veunes · 6h ago
It's probably why workflow agents feel more reliable: they're built around structure, not just raw prediction
jsemrau · 6h ago
Also you have more control points. It's not just a brain a vat.
debuggerson · 3h ago
The more we chat, the more irrelevant details pile up. For example, a small mention early on might get repeated or build on itself, leading to a lot of unnecessary context.
As the conversation continues, it becomes harder for the model to focus on the main point because it gets tangled in all the extra information. Unlike humans, who can intuitively filter out the noise, LLMs struggle to keep track of what’s truly important in longer, more complex exchanges.
SamPatt · 6h ago
I always felt the derision around the term "prompt engineering" was partially due to people overestimating the importance of the initial prompt and underestimating the importance of managing the ongoing context.
You develop a knack for how to steer the models or start a new conversation through experience. The system or initial prompt are important, but nothing will save you if you naively keep a conversation going too long.
veunes · 6h ago
Yeah, totally. Prompt engineering isn't just about crafting the perfect opener, it's more like conversation management. You start to develop a feel for when things are going off the rails and it's time to reset
WhitneyLand · 56m ago
Any reason to not call bullshit on this paper?
One of the biggest developments in language models over the last year has been test-time reasoning (aka inference scaling or “thinking”). Most vendors tested offer such a model. It’s plausible it could make a huge difference here, and they did not bother to test it or even mention it?
Things like COT and planning can really affect this and those are just a couple of things that happen automatically in more advanced models.
Seems like it wouldn’t have been hard to add this to the experiment, but they could’ve called it out in a “Limitations” or “Future Work” section. Or at least a single sentence like “We did not test chain-of-thought prompting, which may mitigate some of these issues”.
dr_dshiv · 8h ago
This is the best paper on machine psychology [1] I’ve yet seen. Rigorous, empirical, insightful — and very practical.
Ha, kind of funny to see this right now. I've been fighting copilot in vscode in trying to get it to output anything once I take the context down to a very specific problem. It feels like I have to reset and almost reground the model into what I'm trying to accomplish at a certain point.
veunes · 6h ago
Kind of wild how even the best models still struggle with keeping context straight over time. Definitely feels like a big challenge if we want these things to hold real conversations.
dontreact · 10h ago
My take: multi turn evals are hard because to do it really correctly you have to simulate a user. This is not yet modeled well enough for multi turn to work as well as it could.
coderatlarge · 9h ago
i’ve see deepseek-coder local get into an infinite loop generating the same line over and over. which i assume without evidence is some sort of feedback from the generated line back into the generation process. so kind of getting lost in thought and going off topic from the simple .h api that my prompt asked for.
silisili · 7h ago
Yes! Deepseek does this to me all the time.
I had 20 something files I wanted it to check and change something. The first 5 or so it did, then the sixth it rightly said everything is correct moving on. It said that for the rest of the 20, the same text over and over.
I checked, and file 6 was the only correct one. It like, learned to just repeat itself after that and did nothing.
TheOtherHobbes · 5h ago
Claude does this too. It gets into a death spiral where it repeats the entire previous output instead of changing parts and moving on.
guardiang · 6h ago
Exactly why expert steering should be valued.
tsunamifury · 9h ago
Have you seen a bunch of humans in a room?
xyzal · 4h ago
I see lots of humans in an echo chamber
alganet · 10h ago
Humans also often get lost in multi-turn conversation.
I have experienced that in person many, many times. Jumps in context that seem easy for one person to follow, but very hard for others.
So, assuming the paper is legit (arxiv, you never know...), its more like something that could be improved than a difference from human beings.
morsecodist · 10h ago
Subjectively the "getting lost" feels totally different than human conversations. Once there is something bad in the context it seems almost impossible to get back on track. All subsequent responses become get a lot worse and it starts contradicting itself. It is possible that with more training this problem can be improved, but what is interesting to me isn't it's worse than humans in this way but that this sort of difficulty scales differently than it does in humans. I would love to get some more objective descriptions of these subjective notions.
alganet · 9h ago
Contradictions are normal. Humans make them all the time. They're even easy to induce, due to the simplistic nature of our communication (lots of ambiguities, semantic disputes, etc).
I don't see how that's a problem.
Subjectivity is part of human communication.
westurner · 9h ago
Algorithmic convergence and caching :: Consensus in conversational human communication
alganet · 9h ago
Any sufficiently large amount of information exchange could be interpreted as computational if you see it as separated parts. It doesn't mean that it is intrinsically computational.
Seeing human interactions as computer-like is a side effect of our most recent shiny toy. In the last century, people saw everything as gears and pulleys. All of these perspectives are essentially the same reductionist thinking, recycled over and over again.
We've seen men promising that they would build a gear-man, resurrect the dead with electricity, and all sorts of (now) crazy talk. People believed it for some time.
imtringued · 2h ago
What you're talking about has absolutely nothing to do with the paper. It's not about jumps in context. It's about LLMs being biased towards producing a complete answer on first try, even when there isn't even enough information. When you provide them with additional information, they will stick with the originally wrong answer. This means that you need to frontload all information in the first prompt and if the LLM messes up, you will have to start from scratch. You can't do that with a human at all. There is no such thing as "single turn conversation" with humans. You can't reset the human to a past state.
At the end of the two weeks, I observed that: The LLM was much less likely to become distracted. Sometimes, I would dump whole forum threads or SO posts into it, when it said "this is not what we are seeing here, because of [earlier context or finding]. I eliminated all dead ends logically and informed it of this (yes, it can help with the reflection, but I had to make the decisions). In the end, I found the cause of my issues.
This somewhat confirms what some user here on HN said a few days ago. LLMs are good at compressing complex information into simple one, but not at expanding simple ideas into complex ones. As long as my input was larger than the output (either complexity or length), I was happy with the results.
I could have done this without the LLM. However, it was helpful in that it stored facts from the outset that I had either forgotten or been unable to retrieve quickly in new contexts. It also made it easier to identify time patterns in large log files, which helped me debug my site-to-site connection. I also optimized many other settings along the way, resolving not only the most problematic issue. This meant, in addition to fixing my problem, I learned quite a bit. The 'state' was only occasionally incorrect about my current parameter settings, but this was always easy to correct. This confirms what others already saw: If you know where you are going and treat it as a tool, it is helpful. However, don't try to offload decisions or let it direct you in the wrong direction.
Overall, 350k Tokens used (about 300k words). Here's a related blog post [1] with my overall path, but not directly corresponding to this specific issue. (please don't recommend wireguard; I am aware of it)
https://g.co/gemini/share/7edf8fa373fe
I’m not saying that your approach is wrong. But most LLM workflows are either brute forcing the solution, or seeking a local minima to be stuck in. It’s like doing thousands of experiments of objects falling to figure out gravity while there’s a physics textbooks nearby.
[0]: https://datatracker.ietf.org/doc/html/rfc1661
That said, I’m building a product - not a PPP driver - so the quicker I can fix the problem and move on, the better.
[0] https://datatracker.ietf.org/doc/html/rfc1331
There’s no way I could fully read that RFC in an hour. And that’s before you even know what reading to focus your attention on, so you’re just being a worse LLM at that point.
I have really learned to mistrust and double check every single line those systems produce. Same for writing code. Everything they produce looks nice and reasonable on the surface but when you dig deaper it falls apart unless it's something very very basic.
I totally agree that LLMs are great at compressing information; I've set up the docs feature in Cursor to index several entire large documentation websites for major libraries and it's able to distill relevant information very quickly.
Sometimes it is good to start new chat or switch to Claude.
And it really helps to be very precise with wording of specification what you want to achieve. Or repeat it sometimes with some added request lines.
GIGO in reality :)
This feels like something that can be fixed with manual instructions which prompt the model to summarize and forget. This might even map appropriately to human psychology. Working Memory vs Narrative/Episodic Memory.
This seems to be in flux now due to RL training on multiturn eval datasets so while the context window is evergreen every time, there will be some bias towards interpreting each prompt as part of a longer conversation. Mutliturn post training is not scaled out yet in public but I think it may be the way to keep on the 'double time spent on goal every 7 months curve'
In many ways this issue could make the Chinese Room thought experiment even more compelling. Because it's a very practical and inescapable issue.
[1] - https://en.wikipedia.org/wiki/Chinese_room
This is mentioned in the Wikipedia page too: "Although its proponents originally presented the argument in reaction to statements of artificial intelligence (AI) researchers, it is not an argument against the goals of mainstream AI research because it does not show a limit in the amount of intelligent behavior a machine can display."
A nice middle-ground I'm finding is to ask Claude an initial conversation starter in its "thinking" mode, and then copy/paste that conversation into LMStudio and have a weaker model like Gemma pick-up from where Claude left off.
Imagine trying to find a specific output/input that was good in the conversation tree.
You can highlight some text in a chat and fork the chat to talk about that text selection, so the LLM has context of that along with the previous chat history and it responds in a new chat (entire chat history up to that point from the parent chat gets copied over - basically inspired by the Unix `fork`).
Your text selection from the parent chat would get turned into a hyperlink to the new child chat so you can always get to it again if you're reading the parent chat.
The fundamental issue is that LLMs do not currently have real long term memory, and until they do, this is about the best we can do.
But it would indeed be nice to either disable answers (without deleting them) or forking a conversation. It wouldn't be hard to implement; I wonder if there's a market for just this?
https://github.com/actualwitch/experiment
Well, someone did something to it; whether it was training, feature boosting the way Golden Gate Claude [0] was done, adjusting the system prompt, or assuring that it's internet search for contextual information would always return material about that, or some combination of those, is neither obvious nor, if someone had a conjecture as to which one or combination it was, easily falsifiable/verifiable.
[0] https://www.anthropic.com/news/golden-gate-claude
Also, let's be honest, in a Musk company they're going to have taking the shortest possible route to accomplishing what he wanted them to.
[0] https://www.cnn.com/2025/05/14/business/grok-ai-chatbot-repl...
Do you have any source on this? System prompts get leaked/extracted all the time so imagine someone would notice this
Edit: just realized you’re talking about the Grok bot, not Grok the LLM available on X or grok.com. With the bot it’s probably harder to extract its exact instructions since it only replies via tweets. For reference here’s the current Grok the LLM system prompt: https://github.com/asgeirtj/system_prompts_leaks/blob/main/g...
As merely 3 of over a dozen examples:
https://x.com/DefiantLs/status/1922213073957327219
https://x.com/PPC4Liberty/status/1922650016579018855
https://x.com/News24/status/1920909178236776755
Context poisoning is not a uniquely LLM problem
Most history would remain, it wouldn’t try to summarize exactly, just prune and organize the history relative to the conversation path?
One could argue that the attention mechanism in transformers is already designed to do that.
But you need to train it more specifically with that in mind if you want it to be better at damping attention to parts that are deemed irrelevant by the subsequent evolution of the conversation.
And that requires the black art of ML training.
While thinking of doing this as a hack on top of the chat product feels more like engineering and we're more familiar with that as a field.
The prompt it uses: https://www.reddit.com/r/ClaudeAI/comments/1jr52qj/here_is_c...
Each time you press enter, you are spinning up a new instance of the LLM and passing in the entire previous chat text plus your new message, and asking it to predict the next tokens. It does this iteratively until the model produces a <stop> token, and then it returns the text to you and the PRODUCT parses it back into separate chat messages and displays it in your UI.
What you are asking the PRODUCT to now do is to edit your and its chat messages in the history of the chat, and then send that as the new history with your latest message. This is the only way to clean the context because the context is nothing more than your messages and its previous responses, plus anything that tools have pulled in. I think it would be sort of a weird feature to add to a chat bot to have the chat bot, each time you send a new message, go back through the entire history of your chat and just start editing the messages to prune out details. You would scroll up and see a different conversation, it would be confusing.
IMO, this is just part of prompt engineering skills to keep your context clean or know how to "clean" it by branching/summarizing conversations.
I have made zed one of my main llm chat interfaces even for non-programming tasks, because being able to do that is great.
It exists in Claude as a true branch - you can see the old threads - and in ChatGPT as without the history.
Edit a previous reply and hit “go” to see it in action.
How often in meetings does everyone maintain a running context of the entire conversation, instead of responding to the last thing that was said with a comment that has an outstanding chance of being forgotten as soon as the next person starts speaking?
No comments yet
This, of course, has certain implications as to the wisdom of the idea of “replacing human programmers”, given that one of the hard parts of the trade is trying to turn vague and often confused ideas into precise specifications by interacting with the shareholders.
IMO the One Weird Trick for LLMs is recognizing that there's no real entity, and that users are being tricked into a suspended-disbelief story.
In most cases cases you're contributing text-lines for a User-character in a movie-script document, and the LLM algorithm is periodically triggered to autocomplete incomplete lines for a Chatbot character.
You can have an interview with a vampire DraculaBot, but that character can only "self-reflect" in the same shallow/fictional way that it can "thirst for blood" or "turn into a cloud of bats."
Badly, and with great difficulty, so while it can just about be done, even then only kinda.
This leads us to new questions: How can we characterize and identify real-world documents which fit? How can we determine what features may be significant, and which of those can be easily transplanted to our use-case?
We see LLM's introspecting all the time[1].
>Notably, DeepSeek-AI et al. report that the average response length and downstreamperformance of DeepSeek-R1-Zero increases as training progresses. They further report an “aha moment” during training, which refers to the “emergence” of the model’s ability to reconsider its previously generated content. As we show in Section 3.2, this reconsideration behaviour is often indicated by the generation of phrases such as ‘wait, ...’ or ‘alternatively, ...’
[1] https://arxiv.org/pdf/2504.07128
I operate LLMs in many conversational modes where it does ask clarifying questions, probing questions, baseline determining questions.
It takes at most one sentence in the prompt to get them to act this way.
What is this one sentence you are using?
I am struggling to elicite clarification behavior form llms
So you can teach a model to sometimes ask for clarification, but will it actually have insight into when it really needs it, or will it just interject for clarification more or less at random? These models have really awful insight into their own capabilities, ChatGPT eg insists to me that it can read braille, and then cheerfully generates a pure hallucination.
That doesn't mean much; humans sometimes do the same thing. I recall a fun story about a mathematician with synesthesia multiplying numbers by mixing the colours together. With a bit of training such a person could also pretend to be executing a normal algorithm for the purposes of passing tests.
I can trivially get any of the foundational models to ask me clarifying questions. I've never had one respond with 'I don't know'.
Which IMO is the name as "idk"
> A token-predictor could still be trained to predict the tokens “I’m not sure what you mean because of points x, y, and z; could you elaborate?”
This is entirely true, and the key insight is even right in your sentence but you don't seem to grasp it. “could still be trained”: you can train an LLM into doing whatever you want it to, but you have to train it specifically for that!
In the beginning of LLM we witnessed this impressive phenomenon where the LLM exhibited emergent capabilities (I'm particularly thinking about LLMs being few shots learners about stuff that wasn't in their training corpus). And these emergent capabilities legitimately raised the question about “how intelligent these things are, really”.
But for the past three years, the key lesson is that this kind of emergent effect is too small to be useful, and the focus has been put towards creating purposely built datasets (with tons of “artificial data”) to train the model to explicitly do things we want it to do. And it works pretty well, as models' capabilities kept improving at a fast pace (and in particular, I don't see would we couldn't overcome the problem highlighted by this paper, with more synthetic data specifically designed for multi-turn conversation). But their progress is now strictly limited by their makers' own intelligence. You cannot just scrap the web throw compute at the problem and expect emergent intelligence to occur anymore. It's more “simulated intelligence” than “artificial intelligence”, really.
Pre-trained LLMs will ask clarifying questions just fine. So I think this is just another consequence of post-training recipes.
Ironically, working with a junior dev is a lot like this -- setting them on a task, then coming back later with dogs and flashlights to retrieve them from the deep woods they've inevitably lost themselves in by just forging ahead, making assumptions, and asking no questions.
Gemini 2.5 Pro and ChatGPT-o3 have often asked me to provide additional details before doing a requested task. Gemini sometimes comes up with multiple options and requests my input before doing the task.
It also mainly happens when the context is clear that we are collaborating on work that will require multiple iterations of review and feedback, like drafting chapters of a handbook.
I have seen ChatGPT ask questions immediately upfront when it relates to medical issues.
The users are being engineered more than the models are, and this isn't the only example.
In the case of medical questions it needs to know further details to provide a relevant diagnosis. That is how it was trained.
In other cases you can observe its reasoning process to see why it would decide to request further details.
I have never seen an LLM just ask questions for the sake of asking. It is always relevant in the context. I don't use them casually. Just wrote a couple of handbooks (~100 pages in a few days). Generating tens of thousands of tokens per session with Gemini.
- "Should I now give you the complete [result], fulfilling [all your demands]?"
- "Just say [go] and I will do it"
- "Do you want either [A, B, or C]"
- "In [5-15] minutes I will give you the complete result"
...
That's an example of what I'm talking about. Watch the reasoning process produce multiple options. That's what it is trained to do. That is problem solving, not "engagement". It requires more compute, not less. You see that more with the expensive models.
> "In [5-15] minutes I will give you the complete result"
I haven't seen that before and I don't see how it's relevant.
Fair point. Thanks for standing your ground and arguing so matter-of-factly with me! Appreciate it.
https://www.reddit.com/r/comics/comments/1l5tbc/update_to_th...
They're great at both tasks, you just have to ask them to do it.
That is, does it actually know when it doesn't know, or are you just making it less confident overall, so it asks questions with no actual insight? Convincing a model to roleplay as someone who doesn't know things vs teaching a model to have insight into when it does and doesn't need clarification seems like a tough one.
We've been working on a lot of data processing and generation tasks. We've been doing this using an API primarily, but sometimes I end up testing creating data in a chat window and I first chat through what the requirements are for the data analysis / processing and then once I'm done I would like the whole conversation to be then summarised into basically a one-prompt process so that I can re-use it (because I can't really process new inputs via the chat).
Even when you do manage to get it down to a single prompt you can use in a chat and then ask the chat to just keep producing new data (like imagine a blog post in certain style if the base content is given as input and I'm making like 20 of them). If you produce these in the chat, there's notable benefits in that if something is wrong with the blog post the chat suggests, you can immediately edit it. The trouble is that the context window starts becoming so big that the chat starts to forget what the original instruction is and eventually you do have to just create a new chat.
One way to solve for this is having a chat with selective memory where you keep a task in memory, but you have the chat forget/not-include all the generated data in the context so that it stays clean, but only bring it to the context if the user refers to it.
Has anyone else done data processing types of tasks in chats and had issues like this? Are there some other tools to use or tricks to do in chats?
https://github.com/t-kalinowski/promptdown
+30pp uplift when using GPT-35-turbo on a mix of 300 tasks.
Free open framework, check the repo try it yourself
https://github.com/AutomationOptimization/tsce_demo
I tested this another 300 times with gpt-4.1 to remove those obtrusive "em-dashes" everyone hates. Tested a single-pass baseline vs TSCE, same exact instructions and prompt "Remove the em-dashes from my linkedin post. . .".
Out of the 300 tests, baseline failed to remove the em-dashes 149/300 times. TSCE failed to remove the em-dashes 18/300 times.
It works, all the data as well as the entire script used for testing is in the repo.
Have you heard of text.replace("—", "-") ?
System prompt: "Remove every em-dash (—) from the following text while leaving other characters unchanged.\n\nReturn only the cleaned text."
User prompt: <prompt from tsce_chat.py filled with em dashes>
Temperature: 0.0
I've built telegram bot http://t.me/experai_bot as univresal UI to LLMs (with somewhat reduced functionality) exactly around idea "non-reply message means new conversation". Wanna keep context? Keep replying to replies of bot. Non-power user strugge with this idea.
--
Also I observed that OpenAI models performed worse replying to the same questions (for example list of options in reply got shorter) even with smallest system message. That was the case with 3.5, 4o. Don't know how modern ones behave. That made me decide not to include any system messages by default Still I give option to add ones if you need. You can even toggle them to mix-and-match.
You can edit responses, sure, but then a bunch of other context is lost.
My flow is basically:
1. plan
2. build
3. branch (into some feature/esoteric dependency issue)
4. goto #2
Prompt pruning/branching should be a first-class tool for any LLM usage.
It dynamically swaps in / out portions of the context. This system is also not based on explicit definitions it relies on LLMs 'filling the gaps'. The system helps the llm break down problems into small tasks which then eventually aggregate into the full task.
In the future such a distinction in memory hierarchies will be more clear
- Primary memory in the training data
- Secondary memory in context
- Tertiary memory in RAG
May I suggest - put what you have out there in the world, even if it’s barely more than a couple of prompts. If people see it and improve on it, and it’s a good idea, it’ll get picked up & worked on by others - might even take on a life of its own!
https://x.com/zacksiri/status/1922500206127349958
You can see it's going from introduction, asking me for my name, and then able to answer question about some topic. There is also another example in the thread you can see.
Behind the scenes, the system prompt is being modified dynamically based on the user's request.
All the information about movies is also being loaded into context dynamically. I'm also working on some technique to unload stuff from context when the subject matter of a given thread has changed dramatically. Imagine having a long thread of conversation with your friend, and along the way you 'context switch' multiple times as time progresses, you probably don't even remember what you said to your friend 4 years ago.
There is a concept of 'main thread' and 'sub threads' involved as well that I'm exploring.
I will be releasing the code base in the coming months. I need to take this demo further than just a few prompt replies.
Is it due to the model's training distribution (mostly single-shot completions), the way context windows are encoded, or an architectural bottleneck?
Feels like there's no dynamic internal state that evolves over the conversation — only a repeated re-parsing of static history. Has anyone seen work on integrating memory/state mechanisms that allow belief revision within a session, not just regurgitation of past tokens?
When you generate future tokens, you're looking at history tokens that are happy.
So how can a model, given sad tokens, generate future happy tokens if it did not learn to do so?
The work you're looking for is already here, it's "thinking". I assume they include sad tokens in the dataset, produce "thinking", which should result in happy tokens coming after thinking tokens. If thinking is bad (by looking at following happy tokens), then it's punished, if good, then descent.
Stuff like this:
1. Do: Best practice for X model is to include at max 10k lines of code + task + CONVENTIONS.md + architecture guidance. Only queue tasks for components that are fairly decoupled from the rest of the codebase (e.g. small modules).
2. Don't: Start a project without a clearly defined architecture in this format. Don't ask for tasks that require X amount of reading hops to understand the logic.
I find it frustrating that companies release their benchmaxxing without helping developers actually use their models. It's more ironic that some people think of these AIs as employees. Employees can work with their boss about the best way to achieve things! With LLMs you don't even know how to communicate with them and as a result their output is unreliable.
I guess chain of thought in theory should do that but having variations on prompt and context might behave differently?
My conclusion was that context needs to be managed well for the LLMs to manage accuracy in replies. Also, it helps to have a planning process ("graph reasoning") before task execution because it guardrails the models thought process.
This also introduces a discussion on general use vs workflow agent implementations as in the former it is much more difficult to generalize all components in structuring effective ReAct patterns.
You develop a knack for how to steer the models or start a new conversation through experience. The system or initial prompt are important, but nothing will save you if you naively keep a conversation going too long.
One of the biggest developments in language models over the last year has been test-time reasoning (aka inference scaling or “thinking”). Most vendors tested offer such a model. It’s plausible it could make a huge difference here, and they did not bother to test it or even mention it?
Things like COT and planning can really affect this and those are just a couple of things that happen automatically in more advanced models.
Seems like it wouldn’t have been hard to add this to the experiment, but they could’ve called it out in a “Limitations” or “Future Work” section. Or at least a single sentence like “We did not test chain-of-thought prompting, which may mitigate some of these issues”.
[1] http://ui.adsabs.harvard.edu/abs/2023arXiv230313988H/abstrac...
I had 20 something files I wanted it to check and change something. The first 5 or so it did, then the sixth it rightly said everything is correct moving on. It said that for the rest of the 20, the same text over and over.
I checked, and file 6 was the only correct one. It like, learned to just repeat itself after that and did nothing.
I have experienced that in person many, many times. Jumps in context that seem easy for one person to follow, but very hard for others.
So, assuming the paper is legit (arxiv, you never know...), its more like something that could be improved than a difference from human beings.
I don't see how that's a problem.
Subjectivity is part of human communication.
Seeing human interactions as computer-like is a side effect of our most recent shiny toy. In the last century, people saw everything as gears and pulleys. All of these perspectives are essentially the same reductionist thinking, recycled over and over again.
We've seen men promising that they would build a gear-man, resurrect the dead with electricity, and all sorts of (now) crazy talk. People believed it for some time.