- Parameters
- 2B
- Quant
- MLX 4-bit
- Size
- 3.0 GB
- Family
- gemma4
tagger-e2b-v1
Rejecteddenseresidentmlx
Outcome as of 20 September 2026, from the seat registry. The write-up below was made when the model was tested.
Why it was tested
A custom tagger, fine-tuned with MLX LoRA on 270 turns, intended to replace the general model in the tagging seat.
What came of it
Cutover cancelled. Scored 74.2% against its own base model's 85.8% on the same runtime, profile and 240 pairs, with non-overlapping intervals. The fine-tune cost 11.6 points and bought nothing. The training data needs revisiting, not the runtime.
Fine-tuned from Gemma 4 E2B.
| Suite | Score | 95% interval | Passed | Runtime · profile | Recorded |
|---|---|---|---|---|---|
| tagger | 77.1% | 71.4–81.9 | Passed 185/240 | llama.cppextraction | 2026-08-23 |
Runs marked unenforced were measured through mlx_lm on a suite that requires constrained decoding. mlx_lm accepts the schema and ignores it, so those numbers do not measure what the suite is designed to measure. They are kept and labelled rather than deleted.
Every run log entry for this model (1)
| Finished | Suite | Runtime · profile | Score | Coverage | Grader · dataset · config |
|---|---|---|---|---|---|
| tagger | llama.cppextraction | 77.1% | Coverage 240/240 | g —d —c — |
- Gemma 4 12BSeated · Tag
12B · dense · q4_0 QAT · 7.0 GB
Tested as a cheap tagger after the 80B proved expensive for a task that runs constantly.
- Gemma 4 E2BMeasured, not seated
2B · dense · q4_0 · 5.0 GB
The base model for the tagger fine-tune, and the smallest thing worth measuring.