- Scored runs
- 9
- Run log entries
- 10
RecordIdentifier
edge0-35b
Every result below is as recorded. Without a written entry there is no stated reason for the test and no verdict.
Recorded runs9 scored · 95% intervals
| Suite | Score | 95% interval | Passed | Runtime · profile | Recorded |
|---|---|---|---|---|---|
| enrichment | 77.8% | 65.1–86.8 | Passed 42/54 | llama.cppextraction | 2026-09-16 |
| epistemics | 97.9% | 89.1–99.6 | Passed 47/48 | llama.cppstudy-nothink | 2026-09-16 |
| grounding | 89.6% | 83.0–93.8 | Passed 112/125 | llama.cppextraction | 2026-09-16 |
| needle | 62.7% | 53.4–71.2 | Passed 69/110 | llama.cppextraction | 2026-09-18 |
| screening | 88.3% | 77.8–94.2 | Passed 53/60 | llama.cppextraction | 2026-09-16 |
| synthesis | 81.7% | 70.1–89.4 | Passed 49/60 | llama.cppextraction | 2026-09-16 |
| tagger | 88.3% | 83.7–91.8 | Passed 212/240 | llama.cppextraction | 2026-09-16 |
| tagger_nullable | 91.7% | 87.5–94.5 | Passed 220/240 | llama.cppextraction | 2026-09-16 |
| worldview | 52.5% | 37.5–67.1 | Passed 21/40 | llama.cppextraction | 2026-09-16 |
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 (10)
| Finished | Suite | Runtime · profile | Score | Coverage | Grader · dataset · config |
|---|---|---|---|---|---|
| needleneedle/1 | not recordedextraction | 62.7% | Coverage 110/110 | g c38dec33aa1c86c2d 5c5199b9eee8a595c cc08042253cc6e13 | |
| tagger_nullabletagger-nullable/1 | not recordedextraction | 91.7% | Coverage 240/240 | g 04e875431e26c4b5d db3addf23a2bc4cec 7b18c32f56523945 | |
| taggertagger/2 | not recordedextraction | 88.3% | Coverage 240/240 | g 63e04d139a0e5ed3d 7ef587bf33e5788ac 7149c1e662a034c9 | |
| taggertagger/2 | not recordedextraction | 85.5% | Coverage 83/83 | g 63e04d139a0e5ed3d 7ef587bf33e5788ac 7149c1e662a034c9 | |
| synthesissynthesis/2 | not recordedextraction | 81.7% | Coverage 60/60 | g 4156dd3c13361e8bd 267d1954e225c5b6c 0955df95170a937b | |
| screeningscreening/1 | not recordedextraction | 88.3% | Coverage 60/60 | g 0661b49a10be2d2ad 64c97ca2269c29afc 60a41f7096ed2f0d | |
| enrichmentenrichment/3 | not recordedextraction | 77.8% | Coverage 54/54 | g b51222c2ca38cfdfd 75361864baa727abc a0c2477ba575e774 | |
| epistemicsepistemics/4 | not recordedstudy-nothink | 97.9% | Coverage 48/48 | g acf84972a0ad5f38d 4bce453fd376553cc 63bebcfa1d0c9c54 | |
| worldviewworldview/2 | not recordedextraction | 52.5% | Coverage 40/40 | g 9acfba1b704feb28d e97835de9e6fc227c c4a49113da942eeb | |
| groundinggrounding/1 | not recordedextraction | 89.6% | Coverage 125/125 | g 30b933dcc6a7e2d4d dcb396faf8b83b0ac eb1910ec33534645 |
Findings4 that name this model
- 2026-09-18edge0 scores 0/8 on worldview's conflicting arm and 81% on a probe that varies one thing at a time; the arm was measuring three skills at once
- 2026-09-18Two seats detect temporal conflicts, not one: k2-horizon-36b matches ornith at 4/4revised
- 2026-09-18Every seat emits valid typed edges; three of five turn a hedged future statement into an asserted fact
- 2026-09-17One baseline was 25 points off and nothing in the stamp could explain it; the other six reproduced exactly