4 ms·

As is noted in the paper from which this is inspired from: GPT-4's image generation capabilities were severely diminished by the instruction/safety-tuning proce

As is noted in the paper from which this is inspired from: GPT-4's image generation capabilities were severely diminished by the instruction/safety-tuning process. Unfortunately this means the currently available model from the API won't be very capable - certainly not as capable as the early version of GPT-4 that Microsoft had access to. edit: I'm specifically referring to the "image generation by trickery (e.g. SVG)" technique being diminished. Other tasks were diminished as well though - is my understanding.

famouswaffles 4y ago
It's not just image generation the rlhf worsens too. Calibration (confidence on solving a question in relation to ability to solve that problem) went from excellent to non existent. and you can see from the report that the base model performed better on a number of tests. Basically a dumber model.
- arthurcolle 4y ago
  Are you saying this specifically for the GPT-4 API endpoint compared to idealized described GPT-4 from the paper?
  - famouswaffles 4y ago
    yes the public api (or on paid chatGPT) vs the base model from the paper
- tbalsam 4y ago
  Not dumber. More biased. Important distinction, especially if we're looking to push back out towards the Pareto Frontier of the problem. RLHF is still very much in its infancy and does not maximize the bias-variance tradeoff by a long shot, in my personal experience.
  - famouswaffles 4y ago
    No dumber. Sure more biased too if you want but also dumber. Open ai have indicated as much.
    - psychphysic 4y ago
      Also generally less creative and insightful. "No I won't do it" becomes a good option no matter what if you turn safety too high.
  - ShamelessC 4y ago
    My understanding is that OpenAI did indeed find diminished capability across a range of tasks after doing RLHF. You're correct to question this though - as I believe the opposite was true of GPT-3 where it improved certain tasks. The benefits from a business perspective were still clear however, and of course the instruction-tuned GPT-4 model still outperformed GPT-3, in general. There are probably some weird edge cases and nuances that I'm missing - and I'd be happy to be corrected.

famouswaffles 4y ago

It's not just image generation the rlhf worsens too. Calibration (confidence on solving a question in relation to ability to solve that problem) went from excellent to non existent. and you can see from the report that the base model performed better on a number of tests. Basically a dumber model.

arthurcolle 4y ago
Are you saying this specifically for the GPT-4 API endpoint compared to idealized described GPT-4 from the paper?
- famouswaffles 4y ago
  yes the public api (or on paid chatGPT) vs the base model from the paper
tbalsam 4y ago
Not dumber. More biased. Important distinction, especially if we're looking to push back out towards the Pareto Frontier of the problem. RLHF is still very much in its infancy and does not maximize the bias-variance tradeoff by a long shot, in my personal experience.
- famouswaffles 4y ago
  No dumber. Sure more biased too if you want but also dumber. Open ai have indicated as much.
  - psychphysic 4y ago
    Also generally less creative and insightful. "No I won't do it" becomes a good option no matter what if you turn safety too high.
- ShamelessC 4y ago
  My understanding is that OpenAI did indeed find diminished capability across a range of tasks after doing RLHF. You're correct to question this though - as I believe the opposite was true of GPT-3 where it improved certain tasks. The benefits from a business perspective were still clear however, and of course the instruction-tuned GPT-4 model still outperformed GPT-3, in general. There are probably some weird edge cases and nuances that I'm missing - and I'd be happy to be corrected.