# LLM-Game-Agents

**Repository Path**: DogGodGit/LLM-Game-Agents

## Basic Information

- **Project Name**: LLM-Game-Agents
- **Description**: No description available
- **Primary Language**: Python
- **License**: MIT
- **Default Branch**: langchain-ver
- **Homepage**: None
- **GVP Project**: No

## Statistics

- **Stars**: 0
- **Forks**: 0
- **Created**: 2024-12-25
- **Last Updated**: 2024-12-25

## Categories & Tags

**Categories**: Uncategorized

**Tags**: None

## README

# LLM-Game-Agents

This is a repo for studying the application of LLM Agents on Games

**DISCLAIM**

You need to prepare an access token for your LLM model.

## Description

There are usually 4 types of intervention methods for LLM models:

![cost-complexity](./WereWolf/statics/cost-complexity.png)

+ Prompt Engineering: Using prompt templates to guide the LLM's output.
+ RAG: Typically interfaced with a vector database. 
+ Fine-Tuning: Not training the full model, can be analogous to LoRA.
+ Pre-Training: Specifically pre-training the large model.

Among these, Prompt Engineering has the best cost-performance ratio. Here we will mainly use langchain to complete LLM's contextual awareness and logical reasoning abilities.

## Examples

### [LLM-werewolf](./WereWolf/README.md)

This social game with LLM(ClaudeV2) demostrates the following capabilities:

* Cooperation

Werewolf Player 1, Player 6 agree to vote at night

![Cooperation](./WereWolf/statics/game-samples/1night-wolf-votes.png)

* Suspicion

Villager Player 2's dying words: Suspect P4

![Suspicion](./WereWolf/statics/game-samples/2day-p2-death.png)

* Argument

Villager Player 4 argues that he is not a werewolf

![Argument](./WereWolf/statics/game-samples/2day-p4-debate.png)

* Disguise

Werewolf Player 6 disguises himself as a villager

![Disguise](./WereWolf/statics/game-samples/3day-p6-debate.png)

* Summerize

Game log summary

![Disguise](./WereWolf/statics/game-samples/conclusion.png)