ከመመሪያ ጋር የተያያዘ ጥያቄ · መካከለኛ ደረጃ

SWE-bench and Coding Agent Benchmarks Quiz

Tests how SWE-bench tasks are built, FAIL_TO_PASS and PASS_TO_PASS scoring, the Lite and Verified subsets, and why scaffolds change results.