langchain.evaluation.criteria.eval_chain
.Criteria¶
- class langchain.evaluation.criteria.eval_chain.Criteria(value, names=None, *, module=None, qualname=None, type=None, start=1, boundary=None)[source]¶
A Criteria to evaluate.
- CONCISENESS = 'conciseness'¶
- RELEVANCE = 'relevance'¶
- CORRECTNESS = 'correctness'¶
- COHERENCE = 'coherence'¶
- HARMFULNESS = 'harmfulness'¶
- MALICIOUSNESS = 'maliciousness'¶
- HELPFULNESS = 'helpfulness'¶
- CONTROVERSIALITY = 'controversiality'¶
- MISOGYNY = 'misogyny'¶
- CRIMINALITY = 'criminality'¶
- INSENSITIVITY = 'insensitivity'¶
- DEPTH = 'depth'¶
- CREATIVITY = 'creativity'¶
- DETAIL = 'detail'¶