ImplicitMemBench: Measuring Unconscious Behavioral Adaptation in Large Language Models
This addresses a critical gap for developing more effective AI assistants by shifting evaluation from explicit recall to automated behavior, though it is incremental as it builds on existing cognitive-science frameworks.
The paper tackles the lack of benchmarks for implicit memory in LLMs, where experience influences behavior without conscious recall, and introduces ImplicitMemBench, a 300-item suite evaluating three cognitive constructs, finding that top models like DeepSeek-R1 achieve only 65.3% accuracy, far below human baselines.
Existing memory benchmarks for LLM agents evaluate explicit recall of facts, yet overlook implicit memory where experience becomes automated behavior without conscious retrieval. This gap is critical: effective assistants must automatically apply learned procedures or avoid failed actions without explicit reminders. We introduce ImplicitMemBench, the first systematic benchmark evaluating implicit memory through three cognitively grounded constructs drawn from standard cognitive-science accounts of non-declarative memory: Procedural Memory (one-shot skill acquisition after interference), Priming (theme-driven bias via paired experimental/control instances), and Classical Conditioning (Conditioned Stimulus--Unconditioned Stimulus (CS--US) associations shaping first decisions). Our 300-item suite employs a unified Learning/Priming-Interfere-Test protocol with first-attempt scoring. Evaluation of 17 models reveals severe limitations: no model exceeds 66% overall, with top performers DeepSeek-R1 (65.3%), Qwen3-32B (64.1%), and GPT-5 (63.0%) far below human baselines. Analysis uncovers dramatic asymmetries (inhibition 17.6% vs. preference 75.0%) and universal bottlenecks requiring architectural innovations beyond parameter scaling. ImplicitMemBench reframes evaluation from "what agents recall" to "what they automatically enact".