# Bench entry 2026-10-03: gpt-oss-20b on an Apple M5 Pro (24GB), second pass with the packaged harness

- FINDINGS.md: the findings, with entry one's values in brackets for the same rows
- PROTOCOL.md, INTAKE.md, LOG.md: what ran, what did not and why, how the product was obtained, the dated log
- evidence/: speed JSON, egress and memory samples, the server log, mandate-runs/ (summary.md, summary.json, run.log, one JSON per run with the transcript and the signed WitnessKit chain), MANIFEST.json and MANIFEST-chain.json (every file hashed and signed), config.json
- bench-gpt-oss-20b-2026-10-02.zip: the same evidence as a bundle, the format a contributed run arrives in
- To check the evidence yourself: download the bundle, unzip it, and run `node bench.mjs verify --evidence <folder>` from the harness package (bench/harness in the harness zip published alongside the entries). It re-hashes every file against the signed manifest and re-scores every run from its recorded calls.

Home-directory paths in the logs are shown as ~ and the account name as "user".