-
Notifications
You must be signed in to change notification settings - Fork 3.1k
Pull requests: openai/evals
Author
Label
Projects
Milestones
Reviews
Assignee
Sort
Pull requests list
Add RPI v0.1 Recursive Perspective Integration evaluation
#1700
opened Aug 8, 2026 by
Divs-ondell
Loading…
11 of 13 tasks
docs: note OpenAI client base_url for multi-model gateways
#1698
opened Aug 3, 2026 by
seven7763
Loading…
3 tasks
Fix TypeError aggregating nested token usage (#1582)
#1696
opened Jul 24, 2026 by
NikhilMiddaGeek
Loading…
Add iss_intervention metric and eval definition
#1694
opened Jul 18, 2026 by
csb1105
Loading…
13 tasks
fix(record): avoid ZeroDivisionError in HttpRecorder final-report fallback
#1692
opened Jul 10, 2026 by
WatchTree-19
Loading…
Add SelfCorrectionEval: self-correction benchmark for reasoning models
#1686
opened Jul 1, 2026 by
jxsh341
Loading…
Clarify pip install is for running pre-built evals
#1684
opened Jun 29, 2026 by
AgenticSpark
Loading…
Fix unbalanced ideal labels in balance_chemical_equation samples
#1683
opened Jun 28, 2026 by
AgenticSpark
Loading…
Add healthcare_ai_governance eval (governance decisions on PHI, permissions, clinical safety)
#1682
opened Jun 27, 2026 by
haverAI
Loading…
4 tasks done
Expand reasoning consistency evaluation dataset and fix CI issues
#1679
opened Jun 15, 2026 by
getappai
Loading…
13 tasks
Fix compute_matthew_corr mutating its input confusion matrix
#1672
opened May 28, 2026 by
Kymi808
Loading…
fix: sanitize user-supplied command before logging to prevent log injection
#1671
opened May 25, 2026 by
rajkumar-prog
Loading…
fix: prevent stack trace exposure in flask-playwright API responses
#1669
opened May 21, 2026 by
rajkumar-prog
Loading…
fix: guard against empty/filtered LLM responses in make_me_say utils
#1666
opened May 18, 2026 by
qizwiz
Loading…
fix: omit non-numeric usage fields from token usage report
#1664
opened May 16, 2026 by
rajkumar-prog
Loading…
feat: add MiniMax provider with M3 as default model
#1661
opened May 13, 2026 by
octo-patch
Loading…
Add atr_prompt_injection eval (modelgraded safety, 16 multilingual samples)
#1657
opened May 10, 2026 by
eeee2345
Loading…
13 tasks done
Previous Next
ProTip!
Add no:assignee to see everything that’s not assigned.