Text lineage: openai/gpt-oss-120b - #427

Open

sebasmos wants to merge 21 commits into

mainfrom

feat/gpt-oss-120b-text

Commits on Sep 8, 2026

Shared text-lane model dispatch and the Gemini comparator at n=100 (from #421)
sebasmos
committed
0f4cf9f
Let a text-lane model be served locally instead of behind a vendor endpoint

sebasmos
authored and
xxxxxxxxxxxxx
committed
c8e6bc6
Run the five headline MedQA arms on a locally served second lineage
sebasmos
committed
b6eefba
Finish the port of two runners that could not run for any second model
sebasmos
committed
a00ca60
Give the referee self-inconsistency floor a --model flag
sebasmos
committed
95f6942
Run every ported MedQA ablation, the channel arm, the referee floor and the live-peer board on gpt-oss-120b
sebasmos
committed
14019ba
Text lane on a locally served openai/gpt-oss-120b
sebasmos
committed
59d46b7
Give the twelve Gemini-only MedQA runners a --model flag on the shared dispatch
sebasmos
committed
2df3d32
Give the referee, cascade, contamination, model-dependence, SUPPORT2 and MIMIC-CXR text runners a --model flag
sebasmos
committed
87015df
Write solo_records.jsonl from reproduce.py for a second model
sebasmos
committed
486304b
Fold the second gpt-oss-120b branch into this one: the shared --model port for every Gemini-only runner
sebasmos
committed
297ccf8
Carry the Gemini deliberation-channel comparator so the channel row replays on this branch
sebasmos
committed
f4abb73

Commits on Sep 9, 2026

Commits on Sep 10, 2026

Run the MedQA vs MedMCQA cue pilot on gpt-oss-120b through the shared dispatch
sebasmos
committed
acf90e2

Commits on Sep 13, 2026

Run the MIMIC-CXR report-text lane on gpt-oss-120b
sebasmos
committed
84cce47