Dual-Confidence Contrastive Decoding for Retrieval-Augmented Generation
Summary
Proposes Dual-Confidence Contrastive Decoding (DCCD), a training-free method for retrieval-augmented generation that handles intra-context conflicts in multi-document settings by combining document-level and token-level confidence signals, and introduces the DRQA benchmark for factual-conflict QA.
View Cached Full Text
Cached at: 07/02/26, 05:37 AM
# Dual-Confidence Contrastive Decoding for Retrieval-Augmented Generation Source: [https://arxiv.org/abs/2607.00570](https://arxiv.org/abs/2607.00570) [View PDF](https://arxiv.org/pdf/2607.00570) > Abstract:Retrieval\-augmented generation \(RAG\) increasingly requires models to answer questions from multiple retrieved documents, where only some sources are relevant and the retrieved bundle may contain stale, noisy, or conflicting evidence\. Existing contrastive decoding methods primarily focus on resolving conflicts between the model's internal memory and the retrieved context\. In contrast, we study the complementary problem of intra\-context conflict in multi\-document RAG\. To evaluate this setting, we introduce DRQA, a factual\-conflict question answering benchmark derived from enterprise deep\-research scenarios, where answers are grounded in synthetic enterprise\-specific facts that are designed not to be recoverable from the model's internal memory\. We further propose Dual\-Confidence Contrastive Decoding \(DCCD\), a training\-free decoding method that combines document\-level confidence, which estimates whether a document appears sufficient for answering the question, with token\-level confidence, which estimates whether that document supports a confident next\-token prediction\. DCCD selects positive and negative document\-conditioned streams using these dual\-confidence signals and scales a document\-level contrast by their confidence margin\. Across DRQA and standard multi\-document QA benchmarks, DCCD achieves the best average performance among full\-context and contrastive decoding baselines, with the largest gains on DRQA\. These results highlight the importance of source\-aware, confidence\-gated decoding when retrieved evidence is internally conflicting\. ## Submission history From: Raymond Li \[[view email](https://arxiv.org/show-email/352905cb/2607.00570)\] **\[v1\]**Wed, 1 Jul 2026 07:55:21 UTC \(114 KB\)
Similar Articles
Decoupled Contrastive Decoding via Expert-Aligned Drafting
This paper introduces Decoupled Contrastive Decoding (DCD), which uses an expert-aligned lightweight proposer for speculative decoding while keeping the contrastive signal only in verification, achieving speedups over vanilla contrastive decoding without degrading output distribution.
From Context-Aware to Conflict-Aware: Generalizing Contrastive Decoding for Knowledge Conflict in LLMs
The paper generalizes contrastive decoding to a conflict-aware paradigm that dynamically allocates authority between external context and parametric priors, proposes the TriState-Bench evaluation protocol, and introduces Adaptive Regime Routing (ARR) to resolve asymmetry between correction and resistance.
Disco-RAG: Discourse-Aware Retrieval-Augmented Generation
Disco-RAG proposes a discourse-aware retrieval-augmented generation framework that integrates discourse signals through intra-chunk discourse trees and inter-chunk rhetorical graphs to improve knowledge synthesis in LLMs. The method achieves state-of-the-art results on QA and summarization benchmarks without fine-tuning.
ConflictRAG: Detecting and Resolving Knowledge Conflicts in Retrieval Augmented Generation
ConflictRAG is a conflict-aware RAG framework that detects, classifies, and resolves knowledge conflicts in retrieved documents, achieving 88.7% detection F1 and 5.3–6.1% correctness gains over baselines while reducing API costs by 62%.
Dynamic Commonsense Coordination for Empathetic Response Generation
Proposes DCC, a dynamic commonsense coordination framework for empathetic response generation that integrates residual-based interaction, association-guided filtering, and iterative decoding, achieving improved emotion classification and response diversity over baselines.