Lambert: Meta Used Rejection Sampling to Bootstrapping Llama 2 RLHF
Nathan Lambert · The Origin and Future of RLHF: the secret ingredient for ChatGPT - with Nathan Lambert · Jan 11, 2024 · at 1:03:01
Dr. Nathan Lambert explains rejection sampling fine-tuning methodology used in Meta's Llama 2 model.
0:00 / 0:22exact quote · 22.3s
▶ Watch the full episode on YouTube → 720p mp4 · rendered on demand · StarZero watermark
“Llama started their RLHF process with this to get some signal out of preference data. That preference data went into a reward model, and then the reward model did a good enough ranking that it was, like, essentially superpowered instruction tuning based on rewards. Works pretty well. Much easier to implement than PPO, because you can use it in all of your kind of like, it's still instruction tuning, so it's the same autoregressive loss.”
quote is from the automated transcript, cleaned for reading: filler sounds and stutters are removed, nothing is rephrased. names can be misheard (the analysis reads context, assessments check outside sources). how →
More from Nathan Lambert
Lambert: Tulu 3 matches or beats Meta Llama 3.1 on core evals
“On, like, core evals for our Suite of models from, I think, eight, seven D and four or five B is based on llama at the time. It's like it matches or beats meta on these core valves.”
Lambert: Deep Research relies on modular RL tasks rather than end-to-end outcomes
“I think the deep research blog post kind of hints that they do a bunch of small scale RL and then poof, the system works. Which I think is much more of what's happening is people train on a bunch of small things and they do some prompting and they see that whe…”
Lambert: RLHF can never be permanently solved
“In the same way that chatbot arena can never be saturated. RLHF can never be solved.”
Lambert: The RL algorithm is not the most important component in reasoning models
“I definitely don't think the algorithm tends to be the most important thing.”
Lambert: Hybrid reasoners may be phased out except for niche uses
“I think in plenty of ways, like hybrid reasoners might just be aged out except for niche applications because quality is so much more important than having a hundred X less inference tokens. It's like you just pay for it and compute and that'll get better.”
Lambert: SFT cannot teach emergent tool use; models must learn via RL environments
“It's very easy to get the model to do tools if you prompt it to, but it's very hard to get the like RL model to learn that the tool is useful. And that's why it's to go through these things where it's like 80 failed tool uses and it still gets it or like it st…”