You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
When this parameter is 1, `TLM.prompt()` simply returns a standard LLM response and does not attempt to auto-improve it.
577
576
This parameter has no effect when `disable_trustworthiness` is True.
578
577
579
-
disable_trustworthiness (bool, default = False): if True, trustworthiness scoring is disabled and TLM will not compute trust scores for responses.
580
-
This is useful when you only want to use custom evaluation criteria or when you want to minimize computational overhead and only need the base LLM response.
581
-
The following parameters will be ignored when `disable_trustworthiness` is True: `num_consistency_samples`, `num_self_reflections`, `num_candidate_responses`, `reasoning_effort`, `similarity_measure`.
582
-
583
-
prompt: The prompt to use for the TLM call. If not provided, the prompt will be
584
-
generated from the messages.
578
+
disable_trustworthiness (bool, default = False): if True, TLM will not compute trust scores,
579
+
useful if you only want to compute custom evaluation criteria.
585
580
586
581
quality_preset: The quality preset to use for the TLM or Trustworthy RAG API.
When this parameter is 1, `TLM.prompt()` simply returns a standard LLM response and does not attempt to auto-improve it.
1155
1148
This parameter has no effect when `disable_trustworthiness` is True.
1156
1149
1157
-
disable_trustworthiness (bool, default = False): if True, trustworthiness scoring is disabled and TLM will not compute trust scores for responses.
1158
-
This is useful when you only want to use custom evaluation criteria or when you want to minimize computational overhead and only need the base LLM response.
1159
-
The following parameters will be ignored when `disable_trustworthiness` is True: `num_consistency_samples`, `num_self_reflections`, `num_candidate_responses`, `reasoning_effort`, `similarity_measure`.
1160
-
1161
-
prompt: The prompt to use for the TLM call. If not provided, the prompt will be
1162
-
generated from the messages.
1150
+
disable_trustworthiness (bool, default = False): if True, TLM will not compute trust scores,
1151
+
useful if you only want to compute custom evaluation criteria.
1163
1152
1164
1153
quality_preset: The quality preset to use for the TLM or Trustworthy RAG API.
0 commit comments