Who Grades the Grader? Co-Evolving Evaluation Metrics and Skills for Self-Improving LLM Agents
Self-evolving agent systems improve by creating, revising, and retiring their own skills, but every such loop rests on a hidden assumption: a reliable evaluation metric already exists. In many real applications it does not. We make three claims. First, metrics can be \emph{evolved}: our metric loop searches compositions of small drawback detectors under a full evolutionary lifecycle, trained to agree with a ten-item anchored reference set, regularized by consensus over unlabeled outputs, and audited against a held-out anchor it never reads, yielding a transparent, inspectable metric rather tha
Lineage graph
Paper → model → repo connections mined from source citations (Tier-1 exact match).
Why these links exist
Every edge carries a method, confidence, and the source snippet that justified it — so bad links are debuggable.
- PossiblePossibly related (embedding) · 61%AgentToolkit/altk-evolve →
- PossiblePossibly related (embedding) · 59%Prism-Shadow/GDPevo →
- PossiblePossibly related (embedding) · 56%Shiyao-Huang/awesome-agent-evolution →
- PossiblePossibly related (embedding) · 53%Giskard-AI/giskard-oss →
- PossiblePossibly related (embedding) · 52%langwatch/langwatch →
- LinkedLinked via arxiv author · 85%Yuanxing Zhang →
“Who Grades the Grader? Co-Evolving Evaluation Metrics and Skills for Self-Improving LLM Agents”
- LinkedLinked via arxiv author · 85%Guanghui Wang →
“Who Grades the Grader? Co-Evolving Evaluation Metrics and Skills for Self-Improving LLM Agents”
- LinkedLinked via arxiv author · 85%Yanwei Cui →
“Who Grades the Grader? Co-Evolving Evaluation Metrics and Skills for Self-Improving LLM Agents”