Back to corpus
experimentexperiment writeup candidatescore 32

Ranker Serving Handoff — ANE/TurboQuant Wraps the Tiny Ranker, Not Gemma

```text anchor ASR hyp + logits/self-score -> deterministic bounded candidates -> frozen logistic candidate ranker -> calibrated mode threshold -> deterministic corrected text ```

Full HTML reader

Read the full artifact

Open in new tab

Extracted abstract or opening context

Gemma is out of the live path. Full-string Gemma failed the latency gate; edit-op Gemma met the schema only under constrained decode and collapsed to `COPY`. The deployable correction model is the tiny deterministic ranker config: That JSON carries feature means/stds, logistic weights/bias, calibrated operating modes, candidate-generator config, and serialized ASR->clean confusion maps. It does not need training rows at inference time. | mode | CER | delta pp | better/same/worse | |---|---:|---:|---:| | baseline | 0.4352 | +0.00 | 0/0/0 | | aggressive/balanced | 0.3986 | -3.66 | 439/41/2 | | conservative | 0.4026 | -3.26 | 381/15/0 | | preservation | 0.4188 | -1.64 | 196/14/0 | Use `conservative` for automatic correction. Use `aggressive`/`balanced` for offline corpus improvement or human-review queues. - `feat_id` - `asr_hyp` - `asr_score` - CTC log-prob access for candidate scoring

Promotion decision

What has to happen next

Attach run IDs, datasets, metrics, and reproduction commands.

Why this is not always a full paper yet

Corpus pages are public-safe readers for discovered workspace artifacts. They are not automatically final papers. A corpus item becomes a polished paper only after the editable source, evidence checkpoints, references, figures, render path, and release status are attached through the paper schema.