TrainLoop @TrainLoop_ai

Post training research trainloop.ai san francisco, ca Joined January 2025

Tweets

13
Followers

571
Following

29
Likes

101

TrainLoop @TrainLoop_ai

2 weeks ago

@nollahealth is pioneering access to care with AI-driven prescriptions. But critical use cases like these require superhuman performance. Together with @nvidia and @nebius, we post-trained the new Nemotron 3.5 VL Super to SoTA on differential diagnosis.

Jackson Stokes @jackson_stokes

2 weeks ago

We partnered with @nvidia, @nollahealth, and @nebiusai to post train Nemotron 3.5 VL Super far past the frontier on dermatology diagnosis. The model set new standards in our internal measures of trainability, and with Nolla it's powering the future of AI in medical care.

3 5 48 8K 11

0 2 5 596 1

View Details

Nolla Health @nollahealth

2 weeks ago

2,500+ sign-ups, 1,200+ attendees, and so much love! 💙 Our Dermalogica x Nolla pop-up took over the West Village with skin scans, skincare, and more! Huge thanks to everyone who came out and @dermalogica for making it happen. More to come soon.

1 3 14 2K 1

View Details

TrainLoop @TrainLoop_ai

3 months ago

TrainLoop is proud to sign Open Weights and American AI Leadership. As our economy transitions to the Autonomy Age, the best organizations will chart their own path, building their intelligence foundation atop open models. aka.ms/OpenLetter

0 1 8 1K 1

View Details

Jackson Stokes @jackson_stokes

6 months ago

We partnered with @mercor 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

6 9 100 18K 88

View Details

TrainLoop @TrainLoop_ai

6 months ago

+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

6 months ago

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 :

3 6 74 7K 97

0 0 4 994 4

View Details

TrainLoop @TrainLoop_ai

6 months ago

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

6 months ago

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 🧵

5 26 251 23K 218

0 0 5 780 0

View Details

TrainLoop @TrainLoop_ai

7 months ago

Specialized models for the most important tasks!

Jackson Stokes @jackson_stokes

7 months ago

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. @nollahealth Part 1 - improving visual reasoning 🧵1/6

6 9 34 5K 3

0 0 2 467 1

View Details

Joan Cabezas @josancamon19

10 months ago

🧵 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.