ContextBench

ContextBench This repository provides: default: the full ContextBench table (single train split). contextbench_verified: a 500-instance subset (single split). Columns The dataset uses a unified schema across sources: instance_id: ContextBench instance id (e.g., SWE-Bench-Verified__python__...). original_inst_id: Origin

Papers using ContextBench (2)

ContextBench: A Benchmark for Context Retrieval in Coding Agents2026Benchmarking LLMs for Fine-Grained Code Review with Enriched Context in Practice2025