undefined

points

by BryceEller1 days ago |

comments

by xtracto1 days ago|

[-]

The problem is that it seems to be an A/B testing thing:

https://news.ycombinator.com/item?id=50422951

So only if you were "selected" for their experiments, your account was nerfed.

by eggdaft1 days ago|

parent|

[-]

For sure, myself and colleague saw this nerfing. He said - new model coming. It was incredibly obvious. I also started getting a lot of "How's Claude doing today" messages at the same time, which is presumably to get feedback on the A/B

by krenel1 days ago|

parent|

[-]

I always skip those. I know it doesn’t matter in the bigger scheme of things, but I don’t want to be classified as a “good feedback provider”.

I’m working on a task. And the task is never to evaluate the agent performance.

by LtdJorge1 days ago|

parent|

[-]

And whenever you press a number, even 0, they keep your full conversation.

by bmh1001 days ago|

parent|

[-]

Source?

by cromka1 days ago|

parent|

prev|

[-]

I strongly believe it will be shit like this that will have people abandon them. They're being way too cocky for a company that pretty much didn't exist 2 years ago

by Tadpole91811 days ago|

parent|

prev|

[-]

But the claim here is "the models are nerfed before a release", not "they A/B test quants / efforts". If it were A/B testing, it doesn't really make sense to correlate it to release windows at all.

by Y_Y1 days ago|

parent|

[-]

The models are nerfed [for some people] before a release

by eggdaft1 days ago|

parent|

prev|

[-]

They might want to see how people react to a new model depending on whether they've been nerfed.

If the nerfing is lower effort (seems to be, presumably why they hide the CoT) then they may be wondering if they can get away permanently with lowering effort on the new release provided it gives a wee improvement.

by yieldcrv1 days ago|

prev|

[-]

Its incredibly easy to roll out a compute instance to X% of users, or route traffic between compute instances

so not everyone will have the same experience. If I was saving on costs that would be the easiest way to do it without deploying, without any nefarious intent. The persons in charge of that likely don't even know about how individual users are impacted

by olalonde1 days ago|

parent|

[-]

At what point does this become false advertising, or even fraud? Imagine paying AWS for a high-performance server, but 5% of customers who order one quietly get a slower, cheaper server and still pay the full price. That's a serious claim and makes me skeptical Anthropic is doing something like that without more evidence.

by kuboble1 days ago|

parent|

[-]

Actually, as a person who rented a dedicated server or two (not from aws) this exact thing happens quite often.

For example a provider quietly underclocks your cpu or puts it is an energy saving plan hoping you won't notice.

by cvwright1 days ago|

parent|

[-]

Or your "dedicated" CPUs on your VM disappear for minutes at a time, or just process a tiny number of instructions

by apt-apt-apt-apt1 days ago|

parent|

prev|

[-]

I think they could easily reduce several parameters to slightly reduce cost (something like "it's the same model, a few less passes, but quality is still 95%+").

Or switch to a faster, cheaper cache mechanism or something that's almost as good, but much cheaper to run, in the name of improving efficiency and keeping up with demand. This switch to cheaper mechanisms is what it feels like to me.

by eggdaft1 days ago|

parent|

prev|

[-]

It's a problem for us. When they nerf the model our productivity really drops (due to the specific use case and process we have). We've started using OpenAI and will gradually use it more and more if this continues.

by cft1 days ago|

parent|

prev|

[-]

The degradation only happens to some accounts but not to all.

https://www.reddit.com/r/ClaudeCode/s/64DXRERLyU

It would be nice to know how they select. I max out my $200 subscription every week and I was selected. On Saturday, the Opus on my account that happened to be supervising a running live process, became unusable.

by tmpz221 days ago|

parent|

[-]

It would be trivial for them to opt-out accounts that are doing benchmarking from a planned regression.

I really really really do not trust companies on the verge of IPOing from goosing their numbers. Even if the lawyers are up their ass, the social pressures are more than enough to compel the behavior. For every good person willing to do the right thing even at the cost of a delayed career progression, there are 20 people willing to do anything not even for a dollar but for an "atta boy!" from the boss.

by Wowfunhappy1 days ago|

parent|

[-]

If it's a private benchmark, how would they know it's a benchmark?

by krackers1 days ago|

parent|

[-]

Look up eval awareness. There are all sorts of subtle signals an LLM could use to know if something is shaped like a benchmark or not. But that's not a bad thing, every animal behaves differently in play vs real-life.

by Wowfunhappy1 days ago|

parent|

[-]

Yes the model can do that, but it would not be "trivial" (as GGP had said) for Anthropic to opt out accounts performing private benchmarks from this theoretical model downgrade people think is happening.

The most plausible approach would be for Anthropic to train some separate classifiers to detect benchmarks. I'm not saying it's completely impossible but IMO this is conspiracy theory levels of effort.

by tmpz221 days ago|

parent|

prev|

[-]

In the context of public relations they would not need to ID every benchmark only the most credible or most public benchmarks many of which are openly available on github and whose problem-set could be fingerprinted.

You're absolutely right that it is a game of whack-a-mole and eventually we may have leading private + credible + popular benchmarks that ultimately dissuade the behavior, if it can be dissuaded.

by cft1 days ago|

parent|

prev|

[-]

That's a major reason to use local models. Or open weights models on a trusted cloud provider. Even if they are way less capable, they're more predictable

by swatcoder1 days ago|

prev|

[-]

"Nerf tracking" is adversarial and exploratory. It's not something Anthropic wants anyone to do, and it's not something that the "nerf tracking" designers are necessarily doing in a foolproof way. Don't assume too much from them either way.

by jstummbillig1 days ago|

parent|

[-]

But do from anecdotal evidence without benchmarks?

by swatcoder1 days ago|

parent|

[-]

Did I suggest that?

Sometimes you just have insufficient evidence to be sure of something either way, and nobody gets to win the argument or shut somebody else down.

In this case, some have reported analytical work[0] that goes beyond "gut feel" anecodotal evidence, but even that's not going to conclusive and may not be trustworthy at all.

[0] https://www.reddit.com/r/ClaudeCode/comments/1xr9x5y/the_ner...

by airesQ1 days ago|

parent|

[-]

For example in that A/B test the user got more effort from the model.

Ok, it probably burns tokens faster, but is it really a nerf?

by jstummbillig1 days ago|

parent|

prev|

[-]

No, you did not comment on it. That is why I asked. Thanks for the clarification!

by rybosworld1 days ago|

prev|

[-]

These nerf trackers do not work. I wish people would stop posting them.

They don't detect a quantized model until the quantization gets down to 4 bits or less.

Humans absolutely do detect quantization well before 4 bits...

Try it yourself if you want. Try a small local model, both quantized and not, and I would be very surprised if it isn't immediately obvious which is which.

by j_maffe1 days ago|

parent|

[-]

But quantized local models also dip in evals.

by rybosworld1 days ago|

parent|

[-]

It doesn't matter if it's local or not, I just suggested local so people could test themselves.

But no, evals and benchmarks do not detect quantization, until around 4 bit quantization.

https://developers.redhat.com/articles/2024/10/17/we-ran-ove...

https://pub.towardsai.net/your-model-passed-safety-evals-the...

by sroussey1 days ago|

parent|

[-]

That is not my experience. I have issues where quantized models fail pretty early so I have evals to catch it.

And the second article seems to reinforce the idea. Yes, broad perplexity tests are not a good proxy for the tests used during training.

Quantizing anything is effectively changing the weights, even if you don’t quantize the weights themselves.

by scrlk1 days ago|

prev|

[-]

Nothing stopping Anthropic (or other labs) from pulling a VW Dieselgate and ensuring that benchmarks are routed to a non-nerfed model.

by swader9991 days ago|

prev|

[-]

Should be fairly routine for anthropic to game a nerf tracker?

by fritzo1 days ago|

prev|

[-]

That must be on anthropic's benchmaxxing CI test

by chinathrow1 days ago|

parent|

[-]

Nerfmaxxing.

It probably need a whistleblower to uncover what they are doing.