Tag
This paper formalizes task-agnostic environment preprocessing for LLM agents, where an agent explores an unfamiliar environment to produce reusable artifacts for a frozen solver, with meta-agent variants achieving high performance across benchmarks.
Last call for paper submissions to the Meta-Agents Workshop at NeurIPS 2026, with the extended deadline closing tomorrow. The workshop focuses on the responsible use of meta-agents and will feature an outstanding speaker and panelist lineup.
The NeurIPS 2026 Workshop on Responsible Use of Meta-Agents issues a call for papers for its December 11/12 event in Sydney, covering topics like automated agent harness design, agentic optimization, and safety, with an all-star speaker lineup.
This paper introduces Shepherd, a functional programming model and runtime substrate for meta-agents that formalizes operations using Lean and records interactions in a Git-like execution trace. It demonstrates significant performance improvements in runtime intervention, counterfactual optimization, and RL training by enabling fast forking and replay of agent states.