Salvato in:
Dettagli Bibliografici
Autori principali: Zhou, Yigeng, Li, Wu, Lu, Yifan, Wang, Yequan, Liu, Xuebo, Wang, Wenya, Yu, Jun, Zhang, Min, Li, Jing
Natura: Preprint
Pubblicazione: 2026
Soggetti:
Accesso online:https://arxiv.org/abs/2605.12185
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866911675362312192
author Zhou, Yigeng
Li, Wu
Lu, Yifan
Wang, Yequan
Liu, Xuebo
Wang, Wenya
Yu, Jun
Zhang, Min
Li, Jing
author_facet Zhou, Yigeng
Li, Wu
Lu, Yifan
Wang, Yequan
Liu, Xuebo
Wang, Wenya
Yu, Jun
Zhang, Min
Li, Jing
contents Large language models accumulate extensive parametric knowledge through pre-training. However, knowledge conflicts occur when outdated or incorrect parametric knowledge conflicts with external knowledge in the context. Existing methods address knowledge conflicts through contrastive decoding, but in conflict-free scenarios, static approaches disrupt output distribution. Other dynamic decoding methods attempt to measure the degree of conflict but still struggle with complex real-world situations. In this paper, we propose a two-stage decoding method called Dynamic Cognitive Reconciliation Decoding (DCRD), to predict and mitigate context-memory conflicts. DCRD first analyzes the attention map to assess context fidelity and predict potential conflicts. Based on this prediction, the input is directed to one of two decoding paths: (1) greedy decoding, or (2) context fidelity-based dynamic decoding. This design enables DCRD to handle conflicts efficiently while maintaining high accuracy and decoding efficiency in conflict-free cases. Additionally, to simulate scenarios with frequent knowledge updates, we constructed ConflictKG, a knowledge conflict QA benchmark. Experiments on four LLMs across six QA datasets show that DCRD outperforms all baselines, achieving state-of-the-art performance.
format Preprint
id arxiv_https___arxiv_org_abs_2605_12185
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Mitigating Context-Memory Conflicts in LLMs through Dynamic Cognitive Reconciliation Decoding
Zhou, Yigeng
Li, Wu
Lu, Yifan
Wang, Yequan
Liu, Xuebo
Wang, Wenya
Yu, Jun
Zhang, Min
Li, Jing
Computation and Language
Artificial Intelligence
Large language models accumulate extensive parametric knowledge through pre-training. However, knowledge conflicts occur when outdated or incorrect parametric knowledge conflicts with external knowledge in the context. Existing methods address knowledge conflicts through contrastive decoding, but in conflict-free scenarios, static approaches disrupt output distribution. Other dynamic decoding methods attempt to measure the degree of conflict but still struggle with complex real-world situations. In this paper, we propose a two-stage decoding method called Dynamic Cognitive Reconciliation Decoding (DCRD), to predict and mitigate context-memory conflicts. DCRD first analyzes the attention map to assess context fidelity and predict potential conflicts. Based on this prediction, the input is directed to one of two decoding paths: (1) greedy decoding, or (2) context fidelity-based dynamic decoding. This design enables DCRD to handle conflicts efficiently while maintaining high accuracy and decoding efficiency in conflict-free cases. Additionally, to simulate scenarios with frequent knowledge updates, we constructed ConflictKG, a knowledge conflict QA benchmark. Experiments on four LLMs across six QA datasets show that DCRD outperforms all baselines, achieving state-of-the-art performance.
title Mitigating Context-Memory Conflicts in LLMs through Dynamic Cognitive Reconciliation Decoding
topic Computation and Language
Artificial Intelligence
url https://arxiv.org/abs/2605.12185