Biased long-term memories in LLM agents cause measurable deviations in tool parameters across 105 scenarios, seven models, and 608 real tools, persisting under standard memory architectures.
Evaluating personalized tool-augmented llms from the perspectives of personalization and proactivity, 2025
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CR 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Memory-Induced Tool-Drift in LLM Agents
Biased long-term memories in LLM agents cause measurable deviations in tool parameters across 105 scenarios, seven models, and 608 real tools, persisting under standard memory architectures.