uqlm.scorers.entropy.SemanticEntropy#

class uqlm.scorers.entropy.SemanticEntropy(llm=None, postprocessor=None, device=None, use_best=True, system_prompt='You are a helpful assistant.', max_calls_per_min=None, use_n_param=False, sampling_temperature=1.0, verbose=False, nli_model_name='microsoft/deberta-large-mnli', max_length=2000, discrete=True)#

Bases: UncertaintyQuantifier

__init__(llm=None, postprocessor=None, device=None, use_best=True, system_prompt='You are a helpful assistant.', max_calls_per_min=None, use_n_param=False, sampling_temperature=1.0, verbose=False, nli_model_name='microsoft/deberta-large-mnli', max_length=2000, discrete=True)#

Class for computing Discrete Semantic Entropy-based confidence scores. For more on semantic entropy, refer to Farquhar et al.(2024) [1].

Parameters:
  • llm (langchain BaseChatModel, default=None) – A langchain llm BaseChatModel. User is responsible for specifying temperature and other relevant parameters to the constructor of their llm object.

  • postprocessor (callable, default=None) – A user-defined function that takes a string input and returns a string. Used for postprocessing outputs.

  • device (str or torch.device input or torch.device object, default="cpu") – Specifies the device that NLI model use for prediction. Only applies to ‘semantic_negentropy’, ‘noncontradiction’ scorers. Pass a torch.device to leverage GPU.

  • use_best (bool, default=True) – Specifies whether to swap the original response for the uncertainty-minimized response based on semantic entropy clusters.

  • system_prompt (str or None, default="You are a helpful assistant.") – Optional argument for user to provide custom system prompt

  • max_calls_per_min (int, default=None) – Specifies how many api calls to make per minute to avoid a rate limit error. By default, no limit is specified.

  • sampling_temperature (float, default=1.0) – The ‘temperature’ parameter for llm model to generate sampled LLM responses. Must be greater than 0.

  • use_n_param (bool, default=False) – Specifies whether to use n parameter for BaseChatModel. Not compatible with all BaseChatModel classes. If used, it speeds up the generation process substantially when num_responses > 1.

  • verbose (bool, default=False) – Specifies whether to print the index of response currently being scored.

  • nli_model_name (str, default="microsoft/deberta-large-mnli") – Specifies which NLI model to use. Must be acceptable input to AutoTokenizer.from_pretrained() and AutoModelForSequenceClassification.from_pretrained()

  • max_length (int, default=2000) – Specifies the maximum allowed string length. Responses longer than this value will be truncated to avoid OutOfMemoryError

Methods

__init__([llm, postprocessor, device, ...])

Class for computing Discrete Semantic Entropy-based confidence scores.

generate_and_score(prompts[, num_responses])

Evaluate discrete semantic entropy score on LLM responses for the provided prompts.

generate_candidate_responses(prompts)

This method generates multiple responses for uncertainty estimation.

generate_original_responses(prompts)

This method generates original responses for uncertainty estimation.

score([responses, sampled_responses])

Evaluate discrete semantic entropy score on LLM responses for the provided prompts.

async generate_and_score(prompts, num_responses=5)#

Evaluate discrete semantic entropy score on LLM responses for the provided prompts.

Return type:

UQResult

Parameters:
  • prompts (list of str) – A list of input prompts for the model.

  • num_responses (int, default=5) – The number of sampled responses used to compute consistency.

Returns:

UQResult, containing data (prompts, responses, and semantic entropy scores) and metadata

Return type:

UQResult

async generate_candidate_responses(prompts)#

This method generates multiple responses for uncertainty estimation. If specified in the child class, all responses are postprocessed using the callable function defined by the user.

Return type:

List[List[str]]

async generate_original_responses(prompts)#

This method generates original responses for uncertainty estimation. If specified in the child class, all responses are postprocessed using the callable function defined by the user.

Return type:

List[str]

score(responses=None, sampled_responses=None)#

Evaluate discrete semantic entropy score on LLM responses for the provided prompts.

Return type:

UQResult

Parameters:
  • responses (list of str, default=None) – A list of model responses for the prompts. If not provided, responses will be generated with the provided LLM.

  • sampled_responses (list of list of str, default=None) – A list of lists of sampled model responses for each prompt. These will be used to compute consistency scores by comparing to the corresponding response from responses. If not provided, sampled_responses will be generated with the provided LLM.

Returns:

UQResult, containing data (responses, sampled responses, and semantic entropy scores) and metadata

Return type:

UQResult

References