haive.games.mafia.aug_llms¶

Specialized LLM configurations for the Mafia game.

This module provides specialized augmented LLM configurations for different aspects of the Mafia game, including:

  • Role-specific analyzer LLMs

  • Strategic decision-making models

  • Game state evaluators

These configurations extend the basic engines.py configurations with more sophisticated models tailored for specific game aspects.

Example

>>> from haive.games.mafia.aug_llms import get_mafia_analyzer
>>>
>>> # Get an analyzer for evaluating player suspicion levels
>>> analyzer = get_mafia_analyzer("suspicion")
>>> analysis = analyzer.invoke(game_state)

Functions¶

get_mafia_analyzer(analyzer_type)

Get a specialized Mafia game analyzer.

Module Contents¶

haive.games.mafia.aug_llms.get_mafia_analyzer(analyzer_type)[source]¶

Get a specialized Mafia game analyzer.

This function returns a configured analyzer LLM for specific Mafia game analysis tasks, such as suspicion evaluation, player psychology, strategy optimization, or voting analysis.

Parameters:

analyzer_type (str) – Type of analyzer to get (“suspicion”, “psychology”, “strategy”, or “voting”)

Returns:

Configured analyzer

Return type:

AugLLMConfig

Raises:

ValueError – If analyzer_type is not recognized

Example

>>> analyzer = get_mafia_analyzer("suspicion")
>>> analysis = analyzer.invoke(game_context)