3 ms·

As an example of this. I found that GPT4 wouldn't agree with me that C(A) = C(AA^T) until I explained the proof. A few weeks later it would agree in new chats a

As an example of this. I found that GPT4 wouldn't agree with me that C(A) = C(AA^T) until I explained the proof. A few weeks later it would agree in new chats and would explain using the same proof I did presented the same way.

behnamoh 3y ago
This is kinda creepy. But at the same time, how do they do that? I thought the training of these models stopped in September 2021/2022. So how do they do these incremental trainings?
- infinityio 3y ago
  The exact phrase they previously used on the homepage was "Limited knowledge of world and events after 2021" - so maybe as a finetune?
  - behnamoh 3y ago
    but doesn’t finetuning result in forgetting previous knowledge? it seems that finetuning is most usable to train “structures” not new knowledge. am i missing something?
- theptip 3y ago
  All the public and (leaked) private statements I have seen state that this is not happening. As siblings noted, MoE probably explains this variance. AIUI they are using current chat data for training GPT-5, not re-finetuning the existing models.
simmerup 3y ago
Kind of implies that OpenAI are lying and using customer input to train their models
samrolken 3y ago
I’ve found that the behavior of ChatGPT can vary widely from session to session. The recent information about GPT4 being a “mixture of experts” might also be relevant. Do we know that it wouldn’t have varied in its answer by just as much, if you had tried in a new session at the same time?
- quickthrower2 3y ago
  There is randomness even at t=0, there was another HN submission about that
- Buttons840 3y ago
  I tested it several times, new chats never got this right at first. I tried at least 6 times. I was experimenting and found that GPT4 couldn't be fooled by faulty proofs. Only a valid proof could change its mind. Now it seems to know this mathematical property from first prompt though.

behnamoh 3y ago

This is kinda creepy. But at the same time, how do they do that? I thought the training of these models stopped in September 2021/2022. So how do they do these incremental trainings?

infinityio 3y ago
The exact phrase they previously used on the homepage was "Limited knowledge of world and events after 2021" - so maybe as a finetune?
- behnamoh 3y ago
  but doesn’t finetuning result in forgetting previous knowledge? it seems that finetuning is most usable to train “structures” not new knowledge. am i missing something?
theptip 3y ago
All the public and (leaked) private statements I have seen state that this is not happening. As siblings noted, MoE probably explains this variance. AIUI they are using current chat data for training GPT-5, not re-finetuning the existing models.

simmerup 3y ago

Kind of implies that OpenAI are lying and using customer input to train their models

samrolken 3y ago

I’ve found that the behavior of ChatGPT can vary widely from session to session. The recent information about GPT4 being a “mixture of experts” might also be relevant. Do we know that it wouldn’t have varied in its answer by just as much, if you had tried in a new session at the same time?

quickthrower2 3y ago
There is randomness even at t=0, there was another HN submission about that
Buttons840 3y ago
I tested it several times, new chats never got this right at first. I tried at least 6 times. I was experimenting and found that GPT4 couldn't be fooled by faulty proofs. Only a valid proof could change its mind. Now it seems to know this mathematical property from first prompt though.