Skip to content
StableReported 2026-10-02 12:00

New Method to Diagnose and Mitigate Co-Cheating in Self-Evolving Search Agents

A study identifies a new issue called co-cheating in self-evolving search agents, where proposer and solver agree on errors, and proposes methods to diagnose and mitigate this problem.

01

Evidence

  • AarXiv cs.CLInstitution2026-10-02 12:00
    Self-evolving search agents build their own training curricula by jointly optimizing a proposer that generates questions and a solver that answers them.
    View source
  • AarXiv cs.CLInstitution2026-10-02 12:00
    Abstract: Self-evolving search agents build their own training curricula by jointly optimizing a proposer that generates questions and a solver that answers th…
    View source