Add reproducible conversational retrieval baselines and paired CIs - #7

Merged

bigboateng merged 12 commits into

mainfrom

cursor/topiocqa-paired-ci-7967

Jul 13, 2026

Commits on Jul 13, 2026

Add paired bootstrap intervals for eval deltas

cursoragent
and
bigboateng
committed
fb9df42
Add pinned TopiOCQA n25 reproduction profile

cursoragent
and
bigboateng
committed
9943ef1
Fix evaluation quality checks

cursoragent
and
bigboateng
committed
ec90842
Format paired evaluation changes

cursoragent
and
bigboateng
committed
7810a19
Avoid repeated embeddings and graph searches in eval

cursoragent
and
bigboateng
committed
60a9d69
Publish reproducible TopiOCQA paired-CI results

cursoragent
and
bigboateng
committed
3170598
Retrigger CI after billing recovery

cursoragent
and
bigboateng
committed
20f5185
Add Gemini query rewrite baseline for CAsT

cursoragent
and
bigboateng
committed
b1de30e
Update Gemini rewrite for current Flash model

cursoragent
and
bigboateng
committed
a5ec616
Fix GitHub Actions version references

cursoragent
and
bigboateng
committed
4df88eb
Format Gemini baseline tests

cursoragent
and
bigboateng
committed
9ec8202
Fix optional dependency CI coverage

cursoragent
and
bigboateng
committed
be9a645