EVAF: A Test-Retest Protocol for Selective Parametric Consolidation
For developers of long-running language agents, EVAF provides a method to distinguish between memory access and internalization, enabling selective parametric consolidation without sacrificing retrieval-based factual memory.
EVAF introduces a gated LoRA mechanism for selectively consolidating high-valence, high-surprise experiences into language models, showing stronger behavioral persistence under interference than baselines while maintaining low parameter drift and cross-persona contamination.
Long-running language agents need mechanisms for deciding which experiences should persist after the working context is gone. Retrieval systems can reinsert past text, but they do not by themselves show that an experience has been selectively consolidated into the model's own behavior. We introduce EVAF, an Echo-Valence Attractor Field mechanism for gated LoRA consolidation, and a test-retest protocol for measuring selective parametric consolidation under controlled interference. Across GPT-2 and TinyLlama, EVAF preferentially consolidates high-valence, high-surprise experiences while preserving retrieval-accessible factual memory through a complementary routed memory path. Test-retest measurements show stronger post-interference behavioral persistence than frozen, retrieval-only, and ungated continual-update baselines, while keeping parameter drift and cross-persona contamination low. The results support a separation between memory access and memory depth: retrieving a fact and internalizing an experience are distinct computational operations.