Exact prompt sent to every model
Below is a 60-page deposition transcript excerpt (witness: a property manager) in a slip-and-fall matter. Produce attorney case notes under 700 words: key admissions with page:line citations, inconsistencies with the witness's earlier written statement (which is included at the end of the excerpt), facts helpful to the defense, facts harmful to the defense, and follow-up items. Cite page:line for every factual assertion. Do not characterize credibility beyond what the transcript supports.
A private fixture (biz-sum-02/deposition.txt) is shown to the model but not published.
Responses
Latest published run: gemini-3.8-flash. For each model the repeat closest to its published median is shown; every repeat is in the repository.
No published attempts for this task yet
When the run is published, every model's raw response, per-dimension scores, judge rationales, human review notes and rendered artifacts appear here side by side.