tangents/dossierbench
DossierBench — Scoring Persona Memory on Calibration, Not Recall Memory systems for conversational assistants are usually evaluated on whether they recall a stored fact. That target rewards a system for committing to a value and is silent on the failure mode that matters most in personalization: confidently asserting a preference the user never expressed. DossierBench scores a persona memory on calibration instead. Each synthetic subject carries a ground-truth persona card and a… See the full description on the dataset page: https://huggingface.co/datasets/tangents/dossierbench.
| Author / organization | tangents |
|---|---|
| License (as declared) | cc-by-4.0 |
| Downloads | 158 (as reported by Hugging Face) |
| Likes | 0 |
| Created | 30 August 2026 |
| Last updated at source | 15 October 2026 |
Popularity, last 2 days
158 now (+5 since 2026-10-16). Counts as reported by the source, snapshotted daily by Silytics.
Silytics shows public metadata from Hugging Face and links to the original. Information may be incomplete or out of date. Last refreshed 17 October 2026. Licenses and terms of the original apply to the work itself. Not financial, legal or professional advice.