1.0.0b3 (2024-10-01)
Features Added
- Added
type
field toAzureOpenAIModelConfiguration
andOpenAIModelConfiguration
- The following evaluators now support
conversation
as an alternative input to their usual single-turn inputs:ViolenceEvaluator
SexualEvaluator
SelfHarmEvaluator
HateUnfairnessEvaluator
ProtectedMaterialEvaluator
IndirectAttackEvaluator
CoherenceEvaluator
RelevanceEvaluator
FluencyEvaluator
GroundednessEvaluator
- Surfaced
RetrievalScoreEvaluator
, formally an internal part ofChatEvaluator
as a standalone conversation-only evaluator.
Breaking Changes
- Removed
ContentSafetyChatEvaluator
andChatEvaluator
- The
evaluator_config
parameter ofevaluate
now maps in evaluator name to a dictionaryEvaluatorConfig
, which is aTypedDict
. The
column_mapping
betweendata
ortarget
and evaluator field names should now be specified inside this new dictionary:
Before:
evaluate(
...,
evaluator_config={
"hate_unfairness": {
"query": "${data.question}",
"response": "${data.answer}",
}
},
...
)
After
evaluate(
...,
evaluator_config={
"hate_unfairness": {
"column_mapping": {
"query": "${data.question}",
"response": "${data.answer}",
}
}
},
...
)
Bugs Fixed
- Fixed issue where Entra ID authentication was not working with
AzureOpenAIModelConfiguration