3 ms·

This is _excellent_ work, I've been adamantly against MoE for a set of reasons, this is the first compelling evidence I've seen that hasn't been on Substack or

This is _excellent_ work, I've been adamantly against MoE for a set of reasons, this is the first compelling evidence I've seen that hasn't been on Substack or a bare repeating of rumor. I had absolutely no idea GPT4 was nondeterministic and I use it about 2 hours a day. I can see why a cursory looking wasn't cutting it, they "feel" the same in your memory, a lot of similar vocab usage, but are formatted entirely differently, and have sort of a synonym-phrase thing going where some of the key words are the same.

derwiki 3y ago
GPT4 web chat for two hours a day? I buy that. Using the API repeatedly for the same inputs, eg developing a program, and the non-determinism is hard to miss.
- sebzim4500 3y ago
  I would imagine that most people use nonzero temperature, so they won't need to look for any explanation for non-determinism.
  - dekhn 3y ago
    Literally the first thing I did when I had llama.cpp working was set the temperature to 0 and repeat queries. (but that's mainly because I'm a weird old scientist with lots of experience with nondeterminism in software).
    - refulgentis 3y ago
      I did too, Kmeans broke me a couple years ago: but, never temperature at 0 with long length, and trusted my instinct instead of actual diffs. This is was the first time I actually diffed
- phillipcarter 3y ago
  Yeah, it's one of the first things you notice when trying to do some kind of "feed GPT some data and get it to produce a novel answer to a question" task with the API.
  - refulgentis 3y ago
    No, because if you wanted a novel answer, why would you set 0 temperature? ;)
152334H 3y ago
Thanks. I'm really no expert (:P) on MoE research; I just noticed what was written in the Soft MoE paper and felt a need to check. The non-deterministic outputs are really similar, yeah, if you check the gist examples I linked https://gist.github.com/152334H/047827ad3740627f4d37826c867a196e https://gist.github.com/152334H/047827ad3740627f4d37826c867a.... This part is at least no surprise, since the randomness should be bounded. I suspect OpenAI will figure out some way to reduce the randomness at some point, though, given their public commitment to eventually adding logprobs back to ChatCompletions.
- cubefox 3y ago
  I don't think this commitment had any plausibility. Token "probabilities" only have a straightforward probabilistic interpretation for base models. In fine-tuned models, they do no longer represent the probability of the next token given the prompt, but rather how well the next token fulfills the ... tendencies induced by SL and RL tuning. Which is presumably pretty useless information. OpenAI has no intention to provide access to the GPT-4 base model, and they in fact removed API access to the GPT-3.5 base model.
  - refulgentis 3y ago
    Topic laundering, the probabilities are the probabilities, you don't suddenly get wrong probabilities with more training on more data
    - goodside 3y ago
      You do, because it’s not just more training it’s PPO updates instead of MLE. It’s no longer trying to estimate the token distribution of the training corpus, it’s trying to shift logprobs into tokens that maximize expected reward from the RM. The GPT-4 technical report has a figure showing that logprobs become less well calibrated as confidence scores in the RLHF vs pre-train model.
      - refulgentis 3y ago
FanaHOVA 3y ago
> I've been adamantly against MoE for a set of reasons Such as?
- lucubratory 3y ago
  It was completely unsubstantiated, based on rumours from a blog, but everyone repeated it as fact.
  - whimsicalism 3y ago
    I think it is pretty compelling that almost all of the people doing research into switch transformers at Google were hired into OAI. I am not sure if that is ouboicly reported but once Ghotz leaked those details about the models, I went to check where the authirs of those papers are now and.... yep
bredren 3y ago
What do you use it for? Are you using many plugins? Curious what sort of insights someone using the tool this much might have, perhaps even through the batch of features released this week.

derwiki 3y ago

GPT4 web chat for two hours a day? I buy that. Using the API repeatedly for the same inputs, eg developing a program, and the non-determinism is hard to miss.

sebzim4500 3y ago
I would imagine that most people use nonzero temperature, so they won't need to look for any explanation for non-determinism.
- dekhn 3y ago
  Literally the first thing I did when I had llama.cpp working was set the temperature to 0 and repeat queries. (but that's mainly because I'm a weird old scientist with lots of experience with nondeterminism in software).
  - refulgentis 3y ago
    I did too, Kmeans broke me a couple years ago: but, never temperature at 0 with long length, and trusted my instinct instead of actual diffs. This is was the first time I actually diffed
phillipcarter 3y ago
Yeah, it's one of the first things you notice when trying to do some kind of "feed GPT some data and get it to produce a novel answer to a question" task with the API.
- refulgentis 3y ago
  No, because if you wanted a novel answer, why would you set 0 temperature? ;)

152334H 3y ago

Thanks. I'm really no expert (:P) on MoE research; I just noticed what was written in the Soft MoE paper and felt a need to check. The non-deterministic outputs are really similar, yeah, if you check the gist examples I linked https://gist.github.com/152334H/047827ad3740627f4d37826c867a196e https://gist.github.com/152334H/047827ad3740627f4d37826c867a.... This part is at least no surprise, since the randomness should be bounded. I suspect OpenAI will figure out some way to reduce the randomness at some point, though, given their public commitment to eventually adding logprobs back to ChatCompletions.

cubefox 3y ago
I don't think this commitment had any plausibility. Token "probabilities" only have a straightforward probabilistic interpretation for base models. In fine-tuned models, they do no longer represent the probability of the next token given the prompt, but rather how well the next token fulfills the ... tendencies induced by SL and RL tuning. Which is presumably pretty useless information. OpenAI has no intention to provide access to the GPT-4 base model, and they in fact removed API access to the GPT-3.5 base model.
- refulgentis 3y ago
  Topic laundering, the probabilities are the probabilities, you don't suddenly get wrong probabilities with more training on more data
  - goodside 3y ago
    You do, because it’s not just more training it’s PPO updates instead of MLE. It’s no longer trying to estimate the token distribution of the training corpus, it’s trying to shift logprobs into tokens that maximize expected reward from the RM. The GPT-4 technical report has a figure showing that logprobs become less well calibrated as confidence scores in the RLHF vs pre-train model.
    - refulgentis 3y ago

FanaHOVA 3y ago

> I've been adamantly against MoE for a set of reasons Such as?

lucubratory 3y ago
It was completely unsubstantiated, based on rumours from a blog, but everyone repeated it as fact.
- whimsicalism 3y ago
  I think it is pretty compelling that almost all of the people doing research into switch transformers at Google were hired into OAI. I am not sure if that is ouboicly reported but once Ghotz leaked those details about the models, I went to check where the authirs of those papers are now and.... yep

bredren 3y ago

What do you use it for? Are you using many plugins? Curious what sort of insights someone using the tool this much might have, perhaps even through the batch of features released this week.