SOURCE-LINKED INTELLIGENCE

Bad Genius: Counterfactual-Guided Harness Evolution Beyond Task-Specific Shortcuts

arXiv · AI, language, vision and robotics · article · Sep 16, 2026 · UTC

Reliable agent evaluation is complicated by automatic harness optimization, which repeatedly uses a released benchmark $B_{\mathrm{rel}}$ to guide a Proposer that edits prompts, memory, retrieval, tools, and control code around a fixed target agent. Task holdout varies semantic tasks but leaves the benchmark protocol fixed, so a "bad genius" Proposer can produce a cheating harness whose released-benchmark gain depends on a benchmark-wide shortcut. We introduce Counterfactual Harness Search and Evolution (CHASE), which casts harness evolution as constraint generation over validity-preserving be

Read original source ↗ Open in workspace

recordType

paper

region

Global