Benchmark
Search-QA
A benchmark that requires an agent to answer questions through multiple rounds of retrieval-based question answering, evaluating information search and integration capabilities.
Timeline (last 90 days)
Oct 15
This reproduction study covers four agent task environments—ALFWorld, ScienceWorld, Search-QA, and WebShop—spans three model scales, and compares 10 representative self-distillation methods against a GRPO baseline.
Unverified50%