What it measures
Wet-lab biological protocol mistake identification benchmark.
Reading the measurement
Ask whether the evaluation rewards recalling a fact, constructing a model, running an experiment or supporting a discovery. Those tasks require different evidence. For research claims, inspect reproducibility, uncertainty and what was actually checked.
Published evaluation
The paper caption explicitly reports these rounded model accuracies, averaged over three runs. It describes other models only as around 7%, so those unspecified individual values are not inferred. Human-expert performance is not a model row.
Where this fits in an evaluation suite
A correct multiple-choice answer and a supported scientific contribution demand different evidence. Check whether the score rewards a proposal, an executed experiment, or a completed research target.
Start with a capability hypothesis, then choose a second test with a different failure mode. Keep the checkpoint, reasoning effort, agent tools, sampling policy, and scoring procedure fixed when measuring a training intervention. Record failures and missing outputs in the denominator.
A reproducible comparison
The paper caption explicitly reports these rounded model accuracies, averaged over three runs. It describes other models only as around 7%, so those unspecified individual values are not inferred. Human-expert performance is not a model row.
Publish the exact run configuration, item coverage, score units, and available uncertainty. If a source supplies only an aggregate or a single run, preserve that limit. A new model name or a rounded score alone does not establish a capability improvement.
PRIMARY EVIDENCE & FURTHER READING
Reviewed 2026-10-27. Explanation and interpretation by RLTasks. Benchmark ownership remains with its original authors.