-
Notifications
You must be signed in to change notification settings - Fork 1.8k
Pull requests: confident-ai/deepeval
Author
Label
Projects
Milestones
Reviews
Assignee
Sort
Pull requests list
fix(metrics): make knowledge retention prompt guardrail-safe
#3034
opened Aug 11, 2026 by
RerankerGuo
Loading…
docs: fix broken internal links after docs restructure
#3031
opened Aug 10, 2026 by
rayyanakmal
Loading…
fix(models): force temperature=1 for o1-preview reasoning models
#3030
opened Aug 10, 2026 by
biztex
Loading…
fix(tool_correctness): give a meaningful reason for reordered repeated tools
#3029
opened Aug 10, 2026 by
Anai-Guo
Contributor
Loading…
fix(metrics): enforce strict_mode in RoleViolationMetric
#3028
opened Aug 10, 2026 by
biztex
Loading…
fix(contextual_recall): strip whitespace before matching verdicts
#3023
opened Aug 9, 2026 by
WatchTree-19
Loading…
feat: report a confidence interval alongside the aggregate pass rate
#3021
opened Aug 9, 2026 by
ipezygj
Loading…
fix: update endpoint and url params for golden methods
#3020
opened Aug 9, 2026 by
A-Vamshi
Collaborator
Loading…
fix(azure): omit temperature for unrecognised deployment names
#3018
opened Aug 8, 2026 by
Aftabbs
Loading…
fix(models): fetch remote multimodal images safely (SSRF + DNS rebinding)
#3012
opened Aug 6, 2026 by
UrielYochpaz
Loading…
fix(benchmarks): gate HumanEval code execution behind explicit opt-in
#3011
opened Aug 6, 2026 by
UrielYochpaz
Loading…
fix(multimodal): don't read local files from untrusted image references
#3010
opened Aug 6, 2026 by
UrielYochpaz
Loading…
fix(prompt): render Jinja templates in a sandbox to prevent SSTI
#3009
opened Aug 6, 2026 by
UrielYochpaz
Loading…
fix(integrations/google_adk): allow offline tracing without CONFIDENT_API_KEY (#3005)
#3007
opened Aug 5, 2026 by
Anai-Guo
Contributor
Loading…
feat: optionally penalize ambiguous answer relevancy verdicts
#3006
opened Aug 5, 2026 by
Yashaswini1233
Loading…
fix: quasi_contains_score doing exact match instead of substring containment
#3004
opened Aug 4, 2026 by
mittalpk
Loading…
fix(g_eval): read the score token from the end of the logprobs (#3000)
#3001
opened Aug 4, 2026 by
Anai-Guo
Contributor
Loading…
fix(optimizer): give each PromptOptimizer its own algorithm instance
#2999
opened Aug 3, 2026 by
shalinis97
Loading…
fix(models): replace nonexistent dated Claude 4.6/4.5 model IDs (default judge 404s)
#2994
opened Aug 2, 2026 by
lokesh75-kank
Loading…
fix: strip <think> tags in trimAndLoadJson for reasoning models
#2991
opened Jul 31, 2026 by
chenrocky
Loading…
Previous Next
ProTip!
Add no:assignee to see everything that’s not assigned.