docs(research): can a local model adjudicate two ASR transcripts?

Eleven local models, 120 calls each, scored against the speaker's blind answers on 30 disagreements, with pass criteria fixed before the run. gemma4-31b (97%) and gemma4-26b (92%) pass; nothing else clears accuracy and order-consistency. Records why 31b is slow (MLX gives the same speed; Low Power Mode is the unproven suspect) and the caveats: 30 items, one recording, thinking models under-budgeted. Aggregate numbers only; this repo is public. Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
2026-10-01 23:14:27 +02:00
parent ebc6cc3ff8
commit e182a6b07d
1 changed files with 75 additions and 0 deletions
@@ -0,0 +1,75 @@
 # Can a local model adjudicate two ASR transcripts? (tapir#33, 2026-10-01)
 Question: when `iguana/kb-whisper` and `iguana/pianissimo` disagree about a word or phrase,
 can a **local** model say which reading is right, well enough to merge the two transcripts?
 Ground truth: the speaker's own blind answers on 30 disagreements of one real recording
 (see `asr-kb-whisper-vs-pianissimo-2026-10-01.md`; no transcript text is recorded here,
 this repo is public). Reference: Claude Opus 5.5 agreed with the speaker on all 30.
 ## Method
 - **Items:** 30 disagreement spans that Opus had picked as changing meaning, the speaker's
  verdict as truth (pianissimo right 17, kb-whisper right 13).
 - **Prompt:** one fixed English instruction (two transcriptions of Swedish speech differ at
  ONE marked spot; say which reading the speaker most likely said; reply `ANSWER: 1|2|UNSURE`
  plus one sentence). Each item was asked in **both reading orders** and in **two contexts**:
  narrow (+-8 words) and wide (+-100 words). 120 calls per model, temperature 0.
 - **Models:** 11, all local (`koala/*`, `iguana/*`): nothing left the homelab. Run one iguana
  model at a time, each unloaded afterwards, to avoid the memory pressure behind infra#273;
  whisper was confirmed healthy afterwards.
 - **Pass criteria, fixed before the run:** wide context, strict accuracy >= 80% (an
  unparseable or "unsure" answer counts as wrong), order-swap consistency >= 85%,
  unparseable <= 10%.
 ## Results (wide context, 60 calls per model)
 | Model | Correct | 95% CI | Same pick when order swapped | Unparseable | s/call | Pass |
 |---|---|---|---|---|---|---|
 | `iguana/gemma4-31b` | 58 = 97% | 89-99% | 93% | 0 | 10.4 | **yes** |
 | `iguana/gemma4-26b` | 55 = 92% | 82-96% | 93% | 0 | 2.0 | **yes** |
 | `iguana/ornith-35b` | 51 = 85% | 74-92% | 77% | 0 | 2.5 | no (consistency) |
 | `koala/phi4-14b` | 50 = 83% | 72-91% | 80% | 0 | 1.0 | no (consistency) |
 | `koala/qwen36-35b-a3b` | 49 = 82% | 70-89% | 83% | 0 | 3.3 | no (consistency) |
 | `iguana/granite4` | 47 = 78% | 66-87% | 83% | 0 | 4.6 | no |
 | `iguana/qwen3-14b` | 44 = 73% | 61-83% | 87% | 0 | 4.4 | no (accuracy) |
 | `iguana/gpt-oss-20b` | 40 = 67% | 54-77% | 100% | 17 | 11.2 | no (unparseable) |
 | `koala/phi4-mini` | 37 = 62% | 49-73% | 24% | 1 | 0.5 | no |
 | `iguana/gemma4-26b-think` | 23 = 38% | 27-51% | 100% | 37 | 15.3 | no (unparseable) |
 | `koala/qwen35-9b-fast` | 0 | n/a | n/a | 60 | 7.2 | no (no output) |
 Narrow context only (+-8 words) costs 6-12 points for the models that pass: gemma4-31b 92%,
 gemma4-26b 83%, ornith-35b 80%. Context is not decoration: Opus also used evidence from
 elsewhere in the recording (an amount stated earlier, a term that recurs).
 Position bias: the passing models pick "Reading 1" 47-51% of the time. `phi4-mini` picks it 15%
 and flips its answer when the order is swapped 76% of the time: it carries almost no signal.
 ## Reading it
 - **Two models pass:** the gemma4 pair. Both clear the bar and neither shows position bias.
  26b is ~5x faster; their intervals overlap, so 30 items cannot separate 92% from 97%.
 - **The thinking models were not tested fairly.** `gpt-oss-20b` and `gemma4-26b-think` spend
  the 600-token budget reasoning and often never reach the answer line. When gemma4-26b-think
  did answer, it was right 23 of 23; gpt-oss-20b 93% of the answered. A larger budget would
  make them eligible; at 15 s a call it would not make them practical.
 ## What this does not show
 - **Thirty items, one recording, one speaker, one domain.** The 30 were chosen by Opus as
  decidable, so they overstate how easy the typical disagreement is; ~350 others are mostly
  spelling, compounds and fillers, which no ground truth covers.
 - Opus agreeing 30/30 is not a measure of Opus over all 380 disagreements.
 - Accuracy was measured at temperature 0 on one prompt; no prompt tuning was tried.
 - The speaker's answers are the only truth; they are one person's recollection of what they said.
 ## Why gemma4-31b is slow, and what was ruled out
 On iguana (M2 Ultra, 64 GB) the dense 31B runs at **103 tok/s prefill, 7.9 tok/s decode**; the
 26B MoE at 641 and 39.6. Ruled out as the cause: memory pressure (no swap, 45% free, model 100%
 on GPU), context length (32k vs 4k made no difference), and the **inference engine** — MLX 4-bit
 of the same model measured 76 tok/s prefill and 8.6 decode, the same within noise. (Terse
 output saves only ~3.6 s of ~9.3 s because prefill dominates.) Two unrelated engines at the same
 speed points at the machine: ~140 GB/s effective memory bandwidth on a part rated for 800, and
 **iguana has macOS Low Power Mode on** (`pmset` `lowpowermode 1`). That is the leading suspect
 and is **not proven**: turning it off needs root and was not done. A first MLX benchmark was
 wrong because `mlx_lm.server` caches prompts; the figures above are from the corrected run.