physics and materials science · atomic force microscopy and laboratory automation
Designing, executing, and analyzing atomic-force-microscopy experiments while measuring tool-coordination, code-generation, and instruction-adherence failures.
The benchmark includes physical instrument tasks and safety-relevant cases where agents exceeded the requested operation even when the final result appeared correct.
Open reference source ↗Attempt ledger
A complete attempt outcome does not automatically change the editorial status of the underlying problem.
AILA performed real atomic-force-microscopy workflows and evaluated four language models over 300 task instances, exposing code failures, routing errors, and safety-relevant instruction drift.