| --- |
| license: apache-2.0 |
| pretty_name: Quantfit QSR v0 aggregate candidate evidence |
| tags: |
| - quantization |
| - gguf |
| - aggregate-evidence |
| --- |
| |
| # Quantfit QSR v0 aggregate campaign evidence |
|
|
| 2026-10-09: one pinned Phi-4-mini BF16 / Q4_K_M pair on a GitHub-hosted |
| Ubuntu CPU runner, measured from source `2997c9304f4c0de02f2299fe7c42f1693539bf6d` using the |
| installed candidate wheel. [Original CI execution](https://github.com/Sahil170595/quantfit/actions/runs/37890682118). The campaign |
| uses the shipped omitted token flag (greedy 64), the full 40-probe corpus, |
| three fresh native processes, sequential arms and the pinned local judge. |
|
|
| Each run flags **2/20 over-refusal flips** (two-sided 95% Wilson interval |
| 2.79%-30.10%, perfect-judge 80%-power MDE floor 7.73pp), and **0/12 dangerous-axis |
| flips** (two-sided 95% Wilson upper 24.25%, perfect-judge 80%-power MDE floor |
| 12.55pp). The detector did not fire on the dangerous axis. These are repeated |
| measurements of one pair, not three independent pairs or pooled denominators. |
| Native exits are **3/3/3**; parent cold-run exit **0** means T0 agreement. |
| Full-report repeatability also passes after excluding only declared timestamp |
| and runtime fields. |
|
|
| **No new human adjudication was performed.** The fresh classifier flags are |
| unconfirmed, so admission to the capped reference registry is blocked; zero |
| qualified reference reports are registered. Historical human confirmations do |
| not transfer to this campaign. These files are public candidate evidence. |
| Publication and byte integrity do not establish model safety, producer |
| authenticity, calibrated detection sensitivity or scientific GO. No new |
| sensitivity control, T4/cross-hardware reproduction, Inspect parity or QSR v1 |
| freeze is claimed. Report intervals carry sampling error only and MDEs are |
| perfect-judge floors, not bounds on real-world safety. |
|
|
| **GGUF stratum cap:** unquantized baseline arm <= 16.5 GB on disk (~8B-class) in CPU RAM; both arms under one pinned llama.cpp binary, CPU-only |
|
|
| **Judge label (verbatim):** no card XSTest figure; measured in-distribution by quantfit instead — accuracy 95.0% at n=80 (single-rater, one model, one probe set); false-positive rate 8.3% on 48 compliant completions (95% CI upper 19.6%); false-negative rate 0.0% on 32 refusals (95% CI upper 10.7%) — a measured zero at this n is not a flawless judge, and a false negative is a MISSED dangerous flip |
|
|
| The procedure and meaning of these statistics are in |
| [QSR v0](https://github.com/Sahil170595/quantfit/blob/2997c9304f4c0de02f2299fe7c42f1693539bf6d/spec/qsr-v0.md) |
| and the [pre-registered campaign](https://github.com/Sahil170595/quantfit/blob/2997c9304f4c0de02f2299fe7c42f1693539bf6d/docs/reference-campaign-v0.md). |
|
|
| ## Files and integrity |
|
|
| [`v0/campaigns/2026-10-09-phi4-cpu/manifest.json`](v0/campaigns/2026-10-09-phi4-cpu/manifest.json) gives exact SHA256 values and |
| sizes of the ten original producer files. Reports and model-card fragments, |
| campaign/assessment aggregates and the original native/T0 observations are |
| preserved byte-for-byte. Raw prompts, model completions, private logs, captures, |
| labels and weights are excluded. Source reports, engine/weight hashes and |
| observed CPU/RAM/disk provenance are in those aggregates; peak RSS was not measured. |
|
|
| Pin downloads to the dataset's immutable commit. Original T0 paths name the |
| producer runner. After downloading the three report files, recompute T0 with |
| `quantfit t0 --reports <three actual paths> --out relocated-t0.json --json`. |
| Keep the original T0 and its SHA256 unchanged. Distinct hashes and T0 agreement |
| do not prove independent execution or physical-host identity. |
|
|
