Termsslug: gated-modular-edits

Gated Modular Edits

门控式改写

Self-improvement strategy that edits one code module at a time and accepts it only if it passes validation and token caps.

Curated Intelligence

1 items · Page 1/1
Rohan Paul (@rohanpaul_ai)T2·Agents
78 pts
Google Study Finds Gated Modular Edits Cut Agent Harness Token Cost by 90%

A new Google study finds that self-improving agent harnesses work best when an optimizer model edits one code module at a time under validation and token caps, rather than performing an unchecked full rewrite. On 40 held-out SWE-bench Lite tasks, the gated approach reached 40.0% accuracy at 46.7k tokens per task, versus 42.5% at 451.6k tokens for a full rewrite, nearly 10x the cost. Enforcing rules via prompts instead of code collapsed accuracy to 2.5%.