Show HN: Mem0 thinks our 2023 conversation happened in 2026

Category: ai-ml

Tags: llm, conversational-memory, benchmark

Score: 5.3/10 (Innovation: 4, Technical: 5, Documentation: 7, Utility: 5)

This project provides the LoCoMo benchmark for evaluating very long-term conversational memory in LLM agents, including synthetic multi-session dialogues with timestamps and annotations for QA and event summarization. It is interesting because it highlights issues in temporal grounding for memory systems, as shown by the comparison with Mem0.

Target audience: AI researchers, ML engineers

Repository: https://www.aurra.us/blog/mem0-vs-aurra · Python · NOASSERTION · 823 stars

View on Hacker News