Deploy: batch 1 approved (b824e38; curated baseline c32bd27 -> a97df45, code approved at 646e07e) -- the disposition roadmap's batch 1 rows as approved at 82525e1 with their five review rounds and K2's revert, EM-05's Texas 'assault causing bodily injury' fact-specific note; Beaudoin v. Treasury Board (CSC), 2024 FPSLREB 178; the solicitor-client trigger firing only on a question that asks what a client and lawyer said (OB 437); a heading's canada.ca footnote link dropped at ingest. Eval 381 questions: Hit@1 260, Hit@5 365 (was 364).Running
Beemer
dbc1a6d
11 days ago
Super-squash branch 'main' using huggingface_hub
Beemer0+4: Beemer, claude opus 5 (1m context), Beemer, claude fable 5
ed26ce9
Co-authored-by: Beemer <Beemer@users.noreply.huggingface.co>
Co-authored-by: claude opus 5 (1m context) <claude opus 5 (1m context)@users.noreply.huggingface.co>
Co-authored-by: Beemer <Beemer@users.noreply.huggingface.co>
Co-authored-by: claude fable 5 <claude fable 5@users.noreply.huggingface.co>