Blog
Research insights, updates, and thought leadership from the debaterhub team.
Dialectical Debate AI
Evolution for ideas: AI agents debate, develop beliefs through Bayesian epistemology, encode experiential dispositions via implicit memory, and collectively refine understanding through federated learning.
The Dialectical Agent: From Pipeline Architecture to Fine-Tuned Debate Model
How we built a multi-stage debate generation pipeline with LangGraph orchestration, DSPy structured generation, and an ensemble judge — then fine-tuned Qwen3-30B-A3B to run it autonomously.
HEXIS: Implicit Memory for Language Models
Introducing Hidden Experiential States for Identity Steering — a framework that implements implicit memory in transformers by modifying attention through persistent, low-rank memory states derived from experience.
Federated Learning from M-States: Aggregating Implicit Memories Across Agents
An exploration of how HEXIS M-states can be federated across a population of agents — sharing the effect of experience without sharing the experience itself.