Love those GPQA scores hovering around 5% when chance (on 4-way multi-choice) would have got them 25%!
montebicyclelo · 18h ago
So could do better than chance by excluding the option it's picked?
gryfft · 19h ago
A stopped clock is right twice a day, but a running clock set to the wrong time is always wrong.
cwt137 · 16h ago
Not always true! Your statement is only true when the running clock's speed is the same as time. Thus, regular time and the clock's time will never meet.
If the clock is running faster than regular time, it will at point catch up to regular time and thus be correct for a split second. If the clock is slower than regular time, regular time will catch up to the clock and the clock will be right for a split second.
k__ · 8h ago
My girlfriend's microwave-clock runs faster than normal.
Somehow this thing manages to accumulate an error of ~15 minutes in a month.
nathan_douglas · 13h ago
If the clock is running backwards at very high speed, it would be right infinitely many times but the proportion of the time that it is right would approach some finite constant.
patapong · 7h ago
And we haven't even touched on the issue of 24-hour format digital clocks, which can at most be right once per day if stopped!
actionfromafar · 15h ago
If we are being pedantic, running clocks never run exactly the same as time. So they'll be right (very) much more seldom than the stopped clock, which is right twice a day.
parrit · 18h ago
The RMS of wrongness of the running clock is probably lower.
nthingtohide · 10h ago
> a running clock set to the wrong time is always wrong.
Could be right within 15 min accuracy in the appropriate timezone. And such a mechanism can be corrected for in the postprocessing step.
dudeinhawaii · 13h ago
or.. A stopped clock is right twice a day; a mis-prompted LLM is wrong 19 times out of 20—but only because we handed it the wrong instruction sheet.
Procedural error in testing perhaps? I'm not familiar with the methodology for GPQA.
anentropic · 6h ago
> they added another trillion tokens and shrank the model from 18 GB to 9 GB through quantization, reducing its bit width from Mamba2’s 16-bit floating-point precision to 8-bits.
This sounds like what they call "Bamba-9B" is actually an 18B model quantised to 8 bits.
I thought generally we were naming models "nB" by their number of params and treating quantisation as a separate concern. Are there any other models that instead treat the name as an indicative memory requirement?
Is this an attempt to hide that it fares poorly vs other ~18B parameter models?
EDIT: no, I just misunderstood
cubefox · 6h ago
> This sounds like what they call "Bamba-9B" is actually an 18B model quantised to 8 bits.
No it doesn't? The fact that it is 18 GB with 16 bit per parameter before quantization means that it is a 9B parameter model.
anentropic · 5h ago
Ah thanks, I see where I got confused now.
tmalsburg2 · 6h ago
Yeah, that's confusing, but the HuggingFace page says it has 9.78 B parameters.
SSM (state space model) -> SSSM (structured state space model) -> (it's like a snake ssss...) Mamba -> Bamba
zaptrem · 10h ago
Where does the B come from?
mentalgear · 19h ago
> chose to make just about everything associated with Bamba open-source — the training recipes, the data, the data loader IBM designed for largescale distributed training, and a quantization framework aimed at shaving storage and inferencing costs.
jmward01 · 21h ago
This type of architecture is definitely the future. Unlimited attn is a dead end. As a human you don't need to scan an entire book just to guess what the next word will be and LLMs shouldn't need that either.
og_kalu · 11h ago
Humans can re-attend to material whenever necessary (i.e you can just re-read a book, re-watch a documentary etc when you feel you have missed crucial context) so it's not the end of the world. These SSMs or modern RNNs can't and if crucial context has been discarded by the end of the query then well too bad. Transformers are of course always re-attending so not an issue for them either. Until that issue is resolved, i don't think attention will be going anywhere.
quantadev · 20h ago
Not be contrarian, but if the next word prediction happens to be someone's name or a place or something discussed multiple places in the book then often, yes, a knowledge of the full plot of the book is "required" just to predict the next word, as you get to the middle or end of a book.
For example you could never fill in the last chapter of any good book without having knowledge of every previous chapter. Not highly detailed knowledge, but still knowledge.
tmalsburg2 · 6h ago
Isn't this exactly the point of this model? No need to memorize everything (which makes transfomers expensive), just keep the relevant info. SSM are essentially recurrent models.
og_kalu · 3h ago
You can't always know what will be "relevant info" in the future. Even humans can't do this but whenever that's an issue, we just go back and re-read, re-watch etc.
None of these modern recurrent architecture have a way to do this.
parrit · 18h ago
What an LLM does is stuff it all into short term memory. Humans dump the first pages into long term memory and "make sense" of it. Humans have a massive context window because of this (and sheer brain size and efficiency).
boroboro4 · 13h ago
We don’t put things into long term memory after we read it. We usually put it after night of sleep. I personally think that context (and kv cache correspondingly) in the models are akin to our short term memory, while training process (and actual weights) are to our long term memory. And we can’t be sure our short term memory doesn’t work in a way of matching the current context towards currently stored short term memory. From this perspective transformers are enough and just fine.
parrit · 13h ago
So if you now hide my original comment and try to recall what I said, do you know it word for word (and are thinking if every word, e.g. did I use one or 2 spaces somewhere as that would change tokens) or do you have a rough concept of what I said?
OTOH if you had to remember a phone number to write it down, how does that differ?
boroboro4 · 12h ago
I think in a way it makes transformers superior to humans, their short term memory is much more powerful =)
Supporting extra long contexts also make transformers super human. Because, again, human's short term memory is exactly this - short term. And much shorter than millions of tokens we expect from models nowadays.
As for SSMs - I think they compress model memory state way too much. Mixed global/local attention layers do just as well. And sparse/block attention seems like a way forward much more (https://arxiv.org/abs/2502.11089).
littlestymaar · 7h ago
> And much shorter than millions of tokens we expect from models nowadays.
Yet all current model still suck above 32k. (Yes some can do needle in a haystack fine, but they still fail at anything even slightly more complex over a long context).
32k is still much higher than humans' though, so I agree with you that it gives them some kind of super human abilities over moderately long context, but they are still disappointingly bad over longer context.
roger_ · 16h ago
Never got how mamba models work in multiple dimensions and non-causally.
Another recent transformer/SSM hybrid is "M1", with a more than 3x claimed inference speed-up compared to equivalent transformers: https://arxiv.org/pdf/2504.10449
IBM is claiming at least a 2x inference speed-up with Bamba. Both groups say that future SSM optimizations to vLLM would lead to further inference speed improvement.
I imported these to America to feed my infant. Data shows the prevalence of peanut allergies lines up with when AAP guidelines started recommending that babies do NOT eat peanut. Israel never went along with this and thus has the lowest rates of allergies in the world.
arijun · 20h ago
I think the difference in allergy rates between UK and Israeli Ashkenazi Jews (10x higher in UK Jews!) [1] is strong evidence for that.
Latest research does strongly suggest that introducing small amounts of common allergens (peanuts, shellfish,milk products...) as early as possible does significantly reduce risk for allergies later. Many early childhood organisations already recommend this. Official health recommendations are often slow to catch up (often for good reasons, but introducing peanuts etc. early is already officially recommended in quite a few countries (Australia, NZ, Sweden for example AFAIK). Not all health professionals are always up to date either though.
itayd · 10h ago
You actually don't need to self import these. Usually Safeway (is it only a west coast thing?) always have these stocked in the Kosher section.
bonzini · 20h ago
As an Italian who has tried (only) the Israeli Bamba, I can certify that it is pretty addictive.
Spot on. From the linked blog post "The refrain of La Bamba, the Mexican folk song that Ritchie Valens made famous, goes: Para bailar La Bamba/Se necesita una poca de Gracia. "
rdtsc · 19h ago
So someone can get fired for picking IBM after all! Or get a bonus, depending on the organization...
fb03 · 15h ago
and in Portuguese, it means "flimsy". What a great name.
folgoris · 17h ago
A very funny and friendly way to say "cocaine" among italians.
I'm struggling to read it seriously.
rzzzt · 22h ago
Para bailar La Bamba / Se necesita una poca de gracia
vienzo · 19h ago
And in Lithuanian it's a navel
dismalaf · 17h ago
Seems like a good fit.
lenerdenator · 17h ago
about time they did something to liven things up at big blue
Love those GPQA scores hovering around 5% when chance (on 4-way multi-choice) would have got them 25%!
If the clock is running faster than regular time, it will at point catch up to regular time and thus be correct for a split second. If the clock is slower than regular time, regular time will catch up to the clock and the clock will be right for a split second.
Somehow this thing manages to accumulate an error of ~15 minutes in a month.
Could be right within 15 min accuracy in the appropriate timezone. And such a mechanism can be corrected for in the postprocessing step.
Procedural error in testing perhaps? I'm not familiar with the methodology for GPQA.
This sounds like what they call "Bamba-9B" is actually an 18B model quantised to 8 bits.
I thought generally we were naming models "nB" by their number of params and treating quantisation as a separate concern. Are there any other models that instead treat the name as an indicative memory requirement?
Is this an attempt to hide that it fares poorly vs other ~18B parameter models?
EDIT: no, I just misunderstood
No it doesn't? The fact that it is 18 GB with 16 bit per parameter before quantization means that it is a 9B parameter model.
https://huggingface.co/ibm-ai-platform/Bamba-9B-fp8
https://en.wikipedia.org/wiki/State-space_representation
Btw bamba if given to kids at a young age can drastically reduce the chance of peanut allergies
https://en.wikipedia.org/wiki/Mitsubishi_Pajero
SSM (state space model) -> SSSM (structured state space model) -> (it's like a snake ssss...) Mamba -> Bamba
For example you could never fill in the last chapter of any good book without having knowledge of every previous chapter. Not highly detailed knowledge, but still knowledge.
None of these modern recurrent architecture have a way to do this.
OTOH if you had to remember a phone number to write it down, how does that differ?
As for SSMs - I think they compress model memory state way too much. Mixed global/local attention layers do just as well. And sparse/block attention seems like a way forward much more (https://arxiv.org/abs/2502.11089).
Yet all current model still suck above 32k. (Yes some can do needle in a haystack fine, but they still fail at anything even slightly more complex over a long context).
32k is still much higher than humans' though, so I agree with you that it gives them some kind of super human abilities over moderately long context, but they are still disappointingly bad over longer context.
IBM is claiming at least a 2x inference speed-up with Bamba. Both groups say that future SSM optimizations to vLLM would lead to further inference speed improvement.
More recently, hybrid architectures that utilize attention plus other operators are gaining traction.
See https://arxiv.org/abs/2503.01868
Also, they sell Bamba at Trader Joe’s now.
[1] https://www.jacionline.org/article/S0091-6749(08)01698-9/ful...
https://en.m.wikipedia.org/wiki/Bamba_(snack)
;)
https://en.wikipedia.org/wiki/La_Bamba_(song)
https://www.thelocal.se/20221125/swedish-word-of-the-day-bam...