4 ms·

Greg Brockman just stated in a long post on Twitter yesterday: "...it’s easy to create a continuum of incrementally-better AIs (such as by deploying subsequent

Greg Brockman just stated in a long post on Twitter yesterday: "...it’s easy to create a continuum of incrementally-better AIs (such as by deploying subsequent checkpoints of a given training run), which presents a safety opportunity very unlike our historical approach of infrequent major model upgrades." [1] This implies OpenAI will be shifting strategy to incrementally releasing future models, so we won't just suddenly see a GPT-5, but GPT-4.x along the way over the next months. The models page you cited also says, "some of our models are now being continually updated," and they will only offer alternative models which are snapshots that are frozen for 3 month windows for those needing more stability in the model. Sebastien Bubeck gave a talk to MIT CSAIL on the Sparks of AGI paper where he commented on how RLHF (training for more safety) has caused the Unicorn test to become less recognizable [2]. He comments about how the safety training is at odds with a lot of these abilities. This seems congruent with the initial results from this project, and it will be interesting to see if they can restore this ability as they continue to push safety training. 1: https://twitter.com/gdb/status/1646183424024268800 2: https://youtu.be/qbIk7-JPB2c?t=1586

imdsm 3y ago
Author here — thanks for explaining this. My expectation is that there will be incremental updates to the model, so while I'm providing the model `gpt-4` for completions, I'm recording the actual model, `gpt-4-0314` in this case, along with the result. I don't want to monitor (and potentially miss) model updates, which is likely as this is very much a fire and forget project that I'll review over time. One per day seems more sensible to me than a large batch per month, as the daily generations are more likely to track model changes if they become frequent. Will be exciting to see where we are in a few months!
- simonw 3y ago
  But if each image is random, what does an image being "better" one day actually tell you about the underlying model? That's why I suggested generating 16 models each time instead of just 1 a- if all 16 are noticeably better than the previous day you've learned something a lot more interesting than if just one appears to be better than the previous one.
  - imdsm 3y ago
    I don't think any individual sample will tell us much. I think we'll need to review this in 3 months, 6 months, 12 months etc, and look for patterns of changes. Ultimately, this is just a fun side project, something to leave running and check back on once every month or two.
  - easygenes 3y ago
    I suggest the author generate a few different examples per day, each with a different temperature setting. 0.0 temperature ought to be deterministic on the same model, but would be interesting to look for trends at higher temperatures over time.
    - imdsm 3y ago
      That's a great idea, thanks easygenes!

imdsm 3y ago

Author here — thanks for explaining this. My expectation is that there will be incremental updates to the model, so while I'm providing the model `gpt-4` for completions, I'm recording the actual model, `gpt-4-0314` in this case, along with the result. I don't want to monitor (and potentially miss) model updates, which is likely as this is very much a fire and forget project that I'll review over time. One per day seems more sensible to me than a large batch per month, as the daily generations are more likely to track model changes if they become frequent. Will be exciting to see where we are in a few months!

simonw 3y ago
But if each image is random, what does an image being "better" one day actually tell you about the underlying model? That's why I suggested generating 16 models each time instead of just 1 a- if all 16 are noticeably better than the previous day you've learned something a lot more interesting than if just one appears to be better than the previous one.
- imdsm 3y ago
  I don't think any individual sample will tell us much. I think we'll need to review this in 3 months, 6 months, 12 months etc, and look for patterns of changes. Ultimately, this is just a fun side project, something to leave running and check back on once every month or two.
- easygenes 3y ago
  I suggest the author generate a few different examples per day, each with a different temperature setting. 0.0 temperature ought to be deterministic on the same model, but would be interesting to look for trends at higher temperatures over time.
  - imdsm 3y ago
    That's a great idea, thanks easygenes!