Skip to content

Latest commit

 

History

10 Commits

Folders and files

NameName
Last commit message
Last commit date
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Agentic AI: Exploring Multi-Agent Debate and Training-Free Self-Improvement

Analyzing reasoning dynamics and self-reflection strategies in large language models.


Overview

This repository is part of an ongoing research study on Agentic AI, focusing on how language models can debate, self-evaluate, and improve their reasoning without explicit fine-tuning.

We explore Training-Free Group Relative Policy Optimization (GRPO) as one of the mechanisms enabling self-improvement through reflection, and compare it to multi-agent debate frameworks where multiple models collaboratively reason to reach consensus.


Core Ideas

Concept Description
Multi-Agent Debate Multiple agents (LLMs) argue for/against different reasoning paths to reach stable, explainable outcomes.
Training-Free GRPO A single-agent variant where the model learns contextually via group-relative feedback, not gradient updates.
Self-Reflection Loops The model summarizes its best reasoning and reuses it as "experience memory" for subsequent prompts.
Reward-Based Comparison Candidates are scored relative to ground truth or heuristic feedback — simulating reinforcement signals.

What This Repository Includes

  • Single-Agent GRPO Framework (training-free)
    → Generates multiple candidates, scores them, and learns via textual reflection.
  • Evaluation Metrics for Agent Reasoning
    → Tracks how reasoning consistency, factual accuracy, and label precision evolve.
  • Audit Traces for Interpretability
    → Stores reasoning steps, rewards, and extracted "experience rules" for transparency.
  • HPC-Ready SLURM Pipelines
    → Fully compatible with UMass Unity GPU nodes for large-scale model experiments.

Why This Matters:

Agentic AI represents the next leap beyond static LLMs — systems that:

  • Argue, reflect, and revise their own reasoning.
  • Adapt to feedback without retraining.
  • Exhibit goal-directed behavior and emergent cooperation.

The GRPO layer acts as a control experiment — testing how far a single agent can self-improve without external debate.

About

Agentic AI - Exploring Multi-Agent Debate and Training-Free Self-Improvement. Experiments with GRPO-inspired feedback loops, reflective memory, and reasoning dynamics in large language models. Part of Independent Study – Agentic AI, UMass Amherst BioNLP Lab.

Resources

Stars

0 stars

Watchers

0 watching

Forks

Releases

Packages

Contributors

Languages