@johnschulman2: People sometimes ask why fine-tune when general-purpose models keep getting better. Bridgewater's work is a good remind…
Summary
John Schulman highlights Bridgewater's work showing fine-tuning with expert-labeled data can significantly outperform prompting-only approaches for financial document sorting.
View Cached Full Text
Cached at: 07/01/26, 01:56 AM
People sometimes ask why fine-tune when general-purpose models keep getting better. Bridgewater’s work is a good reminder that with the right data – here, expert judgements – you can beat prompting-only approaches by a lot. @ddkang and the Bridgewater AIA Labs team are great – glad to see them sharing this.
Tinker (@tinkerapi): Sorting which financial docs are worth an analyst’s time is surprisingly hard for frontier LLMs. With an expert-labeled dataset and on-policy distillation, Bridgewater fine-tuned a model to do it reliably and cheaply.
Similar Articles
@miramurati: Bridgewater used their unique financial knowledge and partnered with us on @tinkerapi to fine-tune a model that helps t…
Bridgewater partnered with Tinker API to fine-tune an LLM to replicate expert judgment in financial tasks, outperforming frontier models on information accuracy and recall at lower cost.
@oneill_c: 1/ We fine-tune a lot of customer models, so we decided to systematically try and figure out some best practices for fi…
The thread shares systematic experimental findings on fine-tuning best practices, varying one SFT lever at a time across dense and MoE models up to 235B on four real-world customer datasets with custom evals to eliminate confounders.
@no_stp_on_snek: Fine Tuning Field Notes Article: short version: it was less about making it smarter and more about giving it a backbone…
A practitioner shares surprising findings from fine-tuning a small open model to be genuinely better in practical use, not just on benchmarks.
Learning to Replicate Expert Judgment in Financial Tasks (14 minute read)
Thinking Machines AI describes training a proprietary LLM to replicate expert judgment in financial document filtering tasks, outperforming frontier models at lower cost.
@no_stp_on_snek: fine-tuning field notes you can give a model better judgment without teaching it anything new. i didn't add knowledge o…
The author shares field notes showing that fine-tuning can improve a model's judgment by steering attention without adding new knowledge or weights, effectively changing its instincts rather than its IQ.