Experiment framework testing how shared vector memory causes recall failures (belief asymmetry, authority collapse, memory anchoring) in multi-agent LLM systems built with LangGraph + ChromaDB (In-Progress).
openai shared-memory multi-agent-systems ai-agents hallucination rag vector-database llm chromadb llm-evaluation langgraph agent-memory agent-simulation memory-anchoring
-
Updated
Jun 22, 2026 - Python