Text-space optimization adapts large language models (LLMs) by editing external natural-language artifacts rather than model weights, keeping the process inspectable and treating the model as a black box. However, most existing methods keep the evaluation fixed, which becomes a bottleneck on open-ended tasks: once the solver improves on the criteria a rubric measures, omitted dimensions remain invisible to the optimization signal.
Simply evolving the rubric is also unreliable when updates are selected by the current solver's score, because apparent progress can come from making the rubric easier to satisfy. To address this, Hugging Face researchers introduce DecoEvo (Decoupled Co-Evolution), which co-evolves a solver skill and a rubric-generator skill under decoupled objectives without using gold rubrics during optimization.
The solver skill is updated using criterion-level feedback, while the rubric-generator skill is revised through complementary audits of requirement coverage and response discrimination that are independent of aggregate solver score. This separation focuses generator updates on newly exposed solver weaknesses, reducing repeated emphasis on criteria the solver already satisfies.
Under each benchmark's official evaluation, DecoEvo outperforms all compared methods across five benchmarks and three LLM backbones, yielding 2.8–5.0% relative gains over SkillOpt in the five-benchmark average. The framework continually refines both problem-solving strategies and evaluation standards by extracting structured feedback from solution audits and analyzing discrepancies across multiple rollouts.