TrainLoop

Post training research

[Jackson Stokes @jackson_stokes]

We partnered with @mercor_ai to test a simple idea:
What if knowledge-work agents were just… coding agents?

Result: +25% performance, 2x faster, cheaper, and new SOTA on APEX-Agents. @josancamon19

[TrainLoop @TrainLoop_ai]

+48% from a single RL step, and 100k rollouts from a single policy.
@LoganGrasby wrote up his findings that off policy training with OAPL is more robust than we thought!

[Jackson Stokes @jackson_stokes]

can we train a model in single RL step?

During recent experiments, @Logangrasby found that a single step of OAPL increased model performance from ~0 to 48% on a clinical reasoning and prediction task.
Turns out, data staleness might matter less than we think. with @pathos :

[TrainLoop @TrainLoop_ai]

Task-specific training can be far more efficient than we realize. This work by @hasith_v explores the LoRA-trained solution space for GSM8k, finding a massive "plane" of solutions, representable in a single rank.

[Jackson Stokes @jackson_stokes]

We trained LoRA adapters of different ranks to understand training dynamics, finding that adapters for GSM8k live in a surprisingly vast, low-rank solution space.

This hints that some model skills are easy to learn, and training is more forgiving than we think. @hasith_v 1/6 🧵

[TrainLoop @TrainLoop_ai]

Specialized models for the most important tasks!

[Jackson Stokes @jackson_stokes]

We post-trained MedGemma to be SoTA in visual medicine ddx, outperforming Opus 4.6, Gemini 3.1 and GPT-5.4 while running at ~1/30th the cost. @getnolla Part 1 - improving visual reasoning 🧵1/6

[Joan Cabezas @josancamon19]

🧵 Labs and VC's are throwing cash at RL environments, especially for computer and browser use. Yet, with just 4 customers and over 30+ vendors, is cloning every website in the world really the path to scale? of course not.

Introducing TRACE: Trajectory Recording and Capture of Environments.

[TrainLoop @TrainLoop_ai]

@FlyaKiet @jackson_stokes @mlpierce22 @avimakesrobots @superset_sh TrainLoop ❤️🤝 Superset

[TrainLoop @TrainLoop_ai]

@jackson_stokes @gdb No way

[TrainLoop @TrainLoop_ai]

@jackson_stokes chatgpt is merging with windsurf is merging with vim

[TrainLoop @TrainLoop_ai]

@elonmusk @EnhancedRadar

[TrainLoop @TrainLoop_ai]

@ThePrimeagen hold my beer