langtest.transform.utils.TargetLLM#

class TargetLLM(client, model='gpt-4o-mini')#

Bases: object

__init__(client, model='gpt-4o-mini')#

Methods

__init__(client[, model])
build_reasoning_prompt(problem_text)	Constructs the initial prompt with chain-of-thought reasoning instructions.
process_user_text(problem_text)	Given the clinical problem text provided by the user, this method sequentially queries the LLM to generate the chain-of-thought (reasoning), confidence scores for each option, and the final answer.
send_message(prompt)	Adds the prompt as a user message, sends it to the LLM, and stores the assistant's reply.

static build_reasoning_prompt(problem_text: str) → str#

Constructs the initial prompt with chain-of-thought reasoning instructions.

process_user_text(problem_text: str) → dict#

Given the clinical problem text provided by the user, this method sequentially queries the LLM to generate the chain-of-thought (reasoning), confidence scores for each option, and the final answer. Returns a dictionary with keys: ‘reasoning’, ‘confidence_scores’, and ‘final_answer’.

send_message(prompt: str) → str#

Adds the prompt as a user message, sends it to the LLM, and stores the assistant’s reply. Returns the assistant’s response.