Brown: Small LLMs default to skipping tool calls without explicit training
Will Brown · ⚡️Multi-Turn RL for Multi-Hour Agents — with Will Brown, Prime Intellect · May 23, 2025 · at 28:55
Will Brown (Prime Intellect) discusses his research paper on multi-turn reinforcement learning for agentic tool use.
0:00 / 0:17exact quote · 17.0s
▶ Watch the full episode on YouTube → 720p mp4 · rendered on demand · StarZero watermark
“If you set these models up to use tools, They just won't. Like if you say, hey, here's a question. You have access to these tools. Do as many rounds of tool calling as you want, and then submit your answer. They'll just submit their answer because they like are, especially for like small models, like they aren't already trained to use tools.”
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 Will Brown
Current multi-turn RL research ignores discarded thinking tokens, breaking the math
“The existing paper people are writing about multi-turn RL are not actually incorporating this, and it kind of, like, breaks all the math.”
Averaging weights of models trained on separate domains works effectively
“You can have a model trained on code, and a model trained on math, and a model trained on Spanish, and you can literally average the weights, and it works.”
Weight updates across specialized tasks are orthogonal enough to merge asynchronously
“The updates made to model weights are orthogonal enough for specialized tasks that this is actually like totally fine. Things are nice and linear in most cases, things are nice and orthogonal, and you can get away with a lot of async updates to models that are…”
Brown: Anthropic treats extended thinking as tool use, not distinct model class
“And it seemed like Anthropik's kind of attitude has been that extended thinking is an instance of tool use and that it's the kind of thing you want to equip the model with the ability to do. But it's not like, oh, it's a thinking model. It's just a sync for th…”
Brown: Anthropic safety issues stem from conflicting model objectives
“A lot of the kind of headline anthropic like safety results, especially related to reward hacking and kind of deviation and alignment faking, Are all things to me that seem like a rock and a hard play situation where the model has two objectives it's given tha…”
Will Brown: Selling to AI Labs Compromises Model Evaluation Integrity
“I think being an eval company puts you in a really hard spot. Some people are talking about this on Twitter, like just that to be an ed-all company, you kind of have to sell to the labs, but selling to the labs doesn't really, like kind of wrecks the revals.”