EnvLunch turns executable environments into free verifiable agent training data
KingsCollegeLondon · hf · 2026-10-28
Researchers from King's College London, Chongqing University and the Alan Turing Institute propose EnvLunch, a warm-up stage that turns executable environments (e.g. databases) into label-free agent training signal: the model samples valid action trajectories, executes them to get final observations, then learns to recover functionally equivalent trajectories from those observations alone. It outperforms the baseline by over 9% on average across downstream RL budgets, and matching its smallest-budget accuracy requires 8-32x more annotated examples. Letting the environment define tasks also beats fine-tuning on same-environment tasks and improves benchmarks beyond data analysis.
More from coding & agent
- Agents hate latency: Claude Opus drops APIs past 7.2s, and 'agent marketing' is born — danshipper · 2026-10-28
- Freelancer lands a Xiaohongshu gig in 5 minutes using Skillry's Motion skill — yihui_indie · 2026-10-28
- LaunchMotion skills let Claude Code render 4K product launch videos in 5 minutes — yihui_indie · 2026-10-28
- Dev rebuilds app with Rork and checks its iPhone Duo listing at Apple's SoHo store — rudrank · 2026-10-28
- Frameworks should be built for AI agents, not humans: the case for rigor over taste — samgoodwin89 · 2026-10-28
- Nous Research ships ElevenLabs SFX plugin for Hermes Agent with looping and variation tools — NousResearch · 2026-10-28