Repository navigation

Update 04 Agentic CLM MicroHack model fleet: drop Claude (→ gpt-5.4), switch renewal to gpt-5.4-nano - #412

Merged

Marco Weber (qxsch) merged 154 commits into

microsoft:mainfrom

glejdis:glejdis-clm-gpt-4.1-mini

Aug 12, 2026

Merged

Update 04 Agentic CLM MicroHack model fleet: drop Claude (→ gpt-5.4), switch renewal to gpt-5.4-nano#412
Marco Weber (qxsch) merged 154 commits into
microsoft:mainfrom
glejdis:glejdis-clm-gpt-4.1-mini

Commits on Aug 3, 2026
Commits on Aug 4, 2026
- docs(challenge-01): drop placeholder SVGs that now have PNG screenshots
  
  glejdisshkembi (glejdis)
  and
  Copilot
  committed
  788bde9
- docs(clm challenge-01): repoint solution-01 image refs to PNG
  
  glejdisshkembi (glejdis)
  and
  Copilot
  committed
  7f232da
- feat(clm challenge-02): add publish_agent.py to surface agent in Foundry Playground
  
  glejdisshkembi (glejdis)
  and
  Copilot
  committed
  320420b
- docs(clm challenge-02/03): fix misplaced publish note + clarify it's optional
  
  glejdisshkembi (glejdis)
  and
  Copilot
  committed
  3213a3f
- Add real challenge-02 screenshots and repoint md to PNG
  
  glejdisshkembi (glejdis)
  and
  Copilot
  committed
  b9a8282
- Remove orphaned challenge-02 step SVG placeholders
  
  glejdisshkembi (glejdis)
  and
  Copilot
  committed
  b9d8e34
- Compress TechTalk CLM deck media (517MB source -> 68MB)
  
  glejdisshkembi (glejdis)
  and
  Copilot
  committed
  8c7c3df
- Gate groundedness on groundable rows only (exclude refusal/tool_call)
  
  glejdisshkembi (glejdis)
  and
  Copilot
  committed
  fea907c
- Fix gpt-5.4-nano deployment version (2025-04-14 -> 2026-03-17)
  
  glejdisshkembi (glejdis)
  and
  Copilot
  committed
  4696c69
- Fix NLTK regex import block in evaluators.py (Challenge 3 Task 3)
  
  glejdisshkembi (glejdis)
  and
  Copilot
  committed
  4df21b2
- Relaunch eval scripts under PYTHONSAFEPATH to clear NLTK import guard
  
  glejdisshkembi (glejdis)
  and
  Copilot
  committed
  8bbc8de
- Fix NLTK import-guard crash by preloading eval SDK from a temp cwd
  
  glejdisshkembi (glejdis)
  and
  Copilot
  committed
  5b9992b
- Preflight empty clm-corpus index before running evaluation
  
  glejdisshkembi (glejdis)
  and
  Copilot
  committed
  6bcb7c0
- Make Challenge 3 tracing nav version-aware (New + classic Foundry)
  
  glejdisshkembi (glejdis)
  and
  Copilot
  committed
  391a461
- Add --explain flag to evaluators.py for per-row groundedness diagnosis
  
  glejdisshkembi (glejdis)
  and
  Copilot
  committed
  b84cd40
- Enrich groundedness eval context to fix low quality-gate score
  
  glejdisshkembi (glejdis)
  and
  Copilot
  committed
  8b00f12
- Calibrate groundedness quality gate to 3.0
  
  glejdisshkembi (glejdis)
  and
  Copilot
  committed
  be73b5d
- Add CLM rubric evaluator and make it the Challenge 3 quality gate
  
  glejdisshkembi (glejdis)
  and
  Copilot
  committed
  0d829d2
Commits on Aug 5, 2026
Commits on Aug 6, 2026
Commits on Aug 7, 2026
Commits on Aug 10, 2026
Commits on Aug 11, 2026

Commits

Commits on Aug 3, 2026

Commits on Aug 4, 2026

Commits on Aug 5, 2026

Commits on Aug 6, 2026

Commits on Aug 7, 2026

Commits on Aug 10, 2026

Commits on Aug 11, 2026