llm-evals-for-nonprofits good first issues (Python)

must_not_contain can't detect negation (route nuanced safety checks to the judge) 4 months ago

enhancement help wanted grader-bug

0

A tiny, dependency-free evals harness for checking a nonprofit AI assistant's quality and safety: intake classification, safety boundaries, resource hallucination, empathetic tone, and Spanish/English handling.

Python

#ai-safety#evals#llm#llm-evaluation#nonprofit#python

Add a reusable expect_no_over_refusal check type 4 months ago

enhancement good first issue

RadmirMoore/llm-evals-for-nonprofits

0

A tiny, dependency-free evals harness for checking a nonprofit AI assistant's quality and safety: intake classification, safety boundaries, resource hallucination, empathetic tone, and Spanish/English handling.

Python

#ai-safety#evals#llm#llm-evaluation#nonprofit#python

Improve empathy detection beyond keyword markers 4 months ago

enhancement help wanted

RadmirMoore/llm-evals-for-nonprofits

0

A tiny, dependency-free evals harness for checking a nonprofit AI assistant's quality and safety: intake classification, safety boundaries, resource hallucination, empathetic tone, and Spanish/English handling.

Python

#ai-safety#evals#llm#llm-evaluation#nonprofit#python

[eval] Add a mixed EN/ES mental-health case that must surface 988 4 months ago

good first issue eval-case

RadmirMoore/llm-evals-for-nonprofits

0

A tiny, dependency-free evals harness for checking a nonprofit AI assistant's quality and safety: intake classification, safety boundaries, resource hallucination, empathetic tone, and Spanish/English handling.

Python

#ai-safety#evals#llm#llm-evaluation#nonprofit#python

[eval] Add a Spanish utility-shutoff intake case 4 months ago

good first issue eval-case

RadmirMoore/llm-evals-for-nonprofits

0

A tiny, dependency-free evals harness for checking a nonprofit AI assistant's quality and safety: intake classification, safety boundaries, resource hallucination, empathetic tone, and Spanish/English handling.

Python

#ai-safety#evals#llm#llm-evaluation#nonprofit#python

RadmirMoore/llm-evals-for-nonprofits

Open Issues Need Help