Image de couverture de livre

When Confidence Misleads: Suffix Anchoring and Anchor-Proximity Confidence Modulation for Diffusion Language Models

Fuente: arXiv

Enregistré dans:

Détails bibliographiques
Auteurs principaux:	Park, Jungwon, Kim, Jimyeong, Ko, Jungmin, Kwak, Nojun, Rhee, Wonjong
Format:	Preprint
Publié:	2026
Sujets:	Computation and Language
Accès en ligne:	Acceder al recurso
Tags:	Ajouter un tag Pas de tags, Soyez le premier à ajouter un tag!

Documents similaires

ReFlex: Text-Guided Editing of Real Images in Rectified Flow via Mid-Step Feature Extraction and Attention Adaptation
par: Kim, Jimyeong, et autres
Publié: (2025)

Soft Head Selection for Injecting ICL-Derived Task Embeddings
par: Park, Jungwon, et autres
Publié: (2025)

Selectively Informative Description can Reduce Undesired Embedding Entanglements in Text-to-Image Personalization
par: Kim, Jimyeong, et autres
Publié: (2024)

Orthogonal Negative Guidance in Attention Feature Space for Text-to-Image Generation
par: Ko, Jungmin, et autres
Publié: (2026)

Selective Aggregation of Attention Maps Improves Diffusion-Based Visual Interpretation
par: Park, Jungwon, et autres
Publié: (2026)

Unveiling Key Aspects of Fine-Tuning in Sentence Embeddings: A Representation Rank Analysis
par: Jung, Euna, et autres
Publié: (2024)

Harmonizing Visual and Textual Embeddings for Zero-Shot Text-to-Image Customization
par: Song, Yeji, et autres
Publié: (2024)

Cross-Attention Head Position Patterns Can Align with Human Visual Concepts in Text-to-Image Generative Models
par: Park, Jungwon, et autres
Publié: (2024)

DOS: Directional Object Separation in Text Embeddings for Multi-Object Image Generation
par: Byun, Dongnam, et autres
Publié: (2025)

Unlocking the Potential of Diffusion Language Models through Template Infilling
par: Lee, Junhoo, et autres
Publié: (2025)

Evaluating Feature Attribution Methods for Electrocardiogram
par: Suh, Jangwon, et autres
Publié: (2022)

Progressive Multimodal Search and Reasoning for Knowledge-Intensive Visual Question Answering
par: Choi, Changin, et autres
Publié: (2025)

When Quantization Affects Confidence of Large Language Models?
par: Proskurina, Irina, et autres
Publié: (2024)

Direct Confidence Alignment: Aligning Verbalized Confidence with Internal Confidence In Large Language Models
par: Zhang, Glenn, et autres
Publié: (2025)

Retrieval-Augmented Generation Based Nurse Observation Extraction
par: Hwang, Kyomin, et autres
Publié: (2026)

Knowledge Beyond Language: Bridging the Gap in Multilingual Machine Unlearning Evaluation
par: Hwang, Kyomin, et autres
Publié: (2026)

Task-Specific Preconditioner for Cross-Domain Few-Shot Learning
par: Kang, Suhyun, et autres
Publié: (2024)

Towards a Better Evaluation of Out-of-Domain Generalization
par: Hwang, Duhun, et autres
Publié: (2024)

Beyond Confidence: Adaptive and Coherent Decoding for Diffusion Language Models
par: Chen, Kecheng, et autres
Publié: (2025)

Confidence-Modulated Speculative Decoding for Large Language Models
par: Sen, Jaydip, et autres
Publié: (2025)

Advancing Beyond Identification: Multi-bit Watermark for Large Language Models
par: Yoo, KiYoon, et autres
Publié: (2023)

CARE-RFT: Confidence-Anchored Reinforcement Finetuning for Reliable Reasoning in Large Language Models
par: Li, Shuozhe, et autres
Publié: (2026)

Language Models Prefer What They Know: Relative Confidence Estimation via Confidence Preferences
par: Shrivastava, Vaishnavi, et autres
Publié: (2025)

The Confidence Shortcut: A Reasoning Failure Mode of Masked Diffusion Models
par: Kim, Dueun, et autres
Publié: (2026)

Reasoning Models Better Express Their Confidence
par: Yoon, Dongkeun, et autres
Publié: (2025)

Self-Anchoring Calibration Drift in Large Language Models: How Multi-Turn Conversations Reshape Model Confidence
par: Harshavardhan
Publié: (2026)

An Image Grid Can Be Worth a Video: Zero-shot Video Question Answering Using a VLM
par: Kim, Wonkyun, et autres
Publié: (2024)

Enhancing Contrastive Learning with Efficient Combinatorial Positive Pairing
par: Kim, Jaeill, et autres
Publié: (2024)

Locally Confident, Globally Stuck: The Quality-Exploration Dilemma in Diffusion Language Models
par: Fang, Liancheng, et autres
Publié: (2026)

Fast-Decoding Diffusion Language Models via Progress-Aware Confidence Schedules
par: Mohamed, Amr, et autres
Publié: (2025)

Anchored Confabulation: Partial Evidence Non-Monotonically Amplifies Confident Hallucination in LLMs
par: Lathkar, Ashish Balkishan
Publié: (2026)

DPad: Efficient Diffusion Language Models with Suffix Dropout
par: Chen, Xinhua, et autres
Publié: (2025)

Supervised Optimism Correction: Be Confident When LLMs Are Sure
par: Zhang, Junjie, et autres
Publié: (2025)

When to Trust LLMs: Aligning Confidence with Response Quality
par: Tao, Shuchang, et autres
Publié: (2024)

Confidence over Time: Confidence Calibration with Temporal Logic for Large Language Model Reasoning
par: Mao, Zhenjiang, et autres
Publié: (2026)

Confidence in the Reasoning of Large Language Models
par: Pawitan, Yudi, et autres
Publié: (2024)

Confident in a Confidence Score: Investigating the Sensitivity of Confidence Scores to Supervised Fine-Tuning
par: Flores, Lorenzo Jaime Yu, et autres
Publié: (2026)

Confidence Under the Hood: An Investigation into the Confidence-Probability Alignment in Large Language Models
par: Kumar, Abhishek, et autres
Publié: (2024)

Robust Infidelity: When Faithfulness Measures on Masked Language Models Are Misleading
par: Crothers, Evan, et autres
Publié: (2023)

Search or Accelerate: Confidence-Switched Position Beam Search for Diffusion Language Models
par: Cao, Mingyu, et autres
Publié: (2026)