Locomo
A long-term memory evaluation benchmark focused on assessing an AI system's cross-session contextual coherence and temporal reasoning capabilities. It requires precise recall of relevant information from large volumes of noisy context.
Timeline (last 90 days)
Locomo (Long-Context Memory) focuses on evaluating cross-session contextual coherence and temporal reasoning
EvoMem is also effective on traditional static benchmarks such as Gaia and Locomo, because multi-turn interactions in long-horizon tasks themselves change the environment state.
All Facts (5)
MemGAS在四个长期记忆基准上进行评估：Locomo、LongMemEval-S、LongMemEval-M和LongMTBench Plus
95%UnverifiedmemU在Locomo测试中达到了92%的准确率，覆盖单跳、多跳、开放域和时序等多种场景
85%UnverifiedLocomo评测数据集存在局限性，不同Memory框架使用不同的judgment prompt导致分数缺乏可比性
80%UnverifiedLocomo (Long-Context Memory) focuses on evaluating cross-session contextual coherence and temporal reasoning
50%UnverifiedEvoMem is also effective on traditional static benchmarks such as Gaia and Locomo, because multi-turn interactions in long-horizon tasks themselves change the environment state.
50%