Meta FAIR Self-Improving Memory Hits 56.0 on Continual Learning Bench
Meta FAIR’s Self-Improving Memory lets Kimi-K2.6 write its own LOAD and OFFLOAD memory prompts, then distills them into its weights. Optimized prompts score 56.0 versus 44.6 for the best hand-written harness on Continual Learning Bench.
Read the full story on MarkTechPost.