Imagen de Portada

Trained on Tokens, Calibrated on Concepts: The Emergence of Semantic Calibration in LLMs

Fuente: arXiv

Guardado en:

Detalles Bibliográficos
Autores principales:	Nakkiran, Preetum, Bradley, Arwen, Goliński, Adam, Ndiaye, Eugene, Kirchhof, Michael, Williamson, Sinead
Formato:	Preprint
Publicado:	2025
Materias:	Computation and Language Machine Learning
Acceso en línea:	Acceder al recurso
Etiquetas:	Agregar Etiqueta Sin Etiquetas, Sea el primero en etiquetar este registro!

Ejemplares similares

Classifier-Free Guidance is a Predictor-Corrector
por: Bradley, Arwen, et al.
Publicado: (2024)

SelfReflect: Can LLMs Communicate Their Internal Answer Distribution?
por: Kirchhof, Michael, et al.
Publicado: (2025)

Step-by-Step Diffusion: An Elementary Tutorial
por: Nakkiran, Preetum, et al.
Publicado: (2024)

BED-LLM: Intelligent Information Gathering with LLMs and Bayesian Experimental Design
por: Choudhury, Deepro, et al.
Publicado: (2025)

Revisiting Uncertainty Quantification Evaluation in Language Models: Spurious Interactions with Response Length Bias Results
por: Santilli, Andrea, et al.
Publicado: (2025)

Vanishing Gradients in Reinforcement Finetuning of Language Models
por: Razin, Noam, et al.
Publicado: (2023)

Annotations Mitigate Post-Training Mode Collapse
por: Springer, Jacob Mitchell, et al.
Publicado: (2026)

Mechanisms of Projective Composition of Diffusion Models
por: Bradley, Arwen, et al.
Publicado: (2025)

Composition and Control with Distilled Energy Diffusion Models and Sequential Monte Carlo
por: Thornton, James, et al.
Publicado: (2025)

Trace Length is a Simple Uncertainty Signal in Reasoning Models
por: Devic, Siddartha, et al.
Publicado: (2025)

The Geometries of Truth Are Orthogonal Across Tasks
por: Azizian, Waiss, et al.
Publicado: (2025)

Uncertainty Quantification for LLM Function-Calling
por: Ye, Zihuiwen, et al.
Publicado: (2026)

Calibration Across Layers: Understanding Calibration Evolution in LLMs
por: Joshi, Abhinav, et al.
Publicado: (2025)

On the Calibration of Multilingual Question Answering LLMs
por: Yang, Yahan, et al.
Publicado: (2023)

Task-Aware Calibration: Provably Optimal Decoding in LLMs
por: Tomov, Tim, et al.
Publicado: (2026)

LLMs are not (consistently) Bayesian: Quantifying internal (in)consistencies of LLMs' probabilistic beliefs
por: Chen, Chacha, et al.
Publicado: (2026)

Refusal Tokens: A Simple Way to Calibrate Refusals in Large Language Models
por: Jain, Neel, et al.
Publicado: (2024)

Beyond the Score: Uncertainty-Calibrated LLMs for Automated Essay Assessment
por: Karim, Ahmed, et al.
Publicado: (2025)

Jailbreaking LLMs via Calibration
por: Lu, Yuxuan, et al.
Publicado: (2026)

Why Semantic Entropy Fails: Geometry-Aware and Calibrated Uncertainty for Policy Optimization
por: Zhang, Zheyuan, et al.
Publicado: (2026)

Token-Driven GammaTune: Adaptive Calibration for Enhanced Speculative Decoding
por: Gautam, Aayush, et al.
Publicado: (2025)

Calibrating Expressions of Certainty
por: Wang, Peiqi, et al.
Publicado: (2024)

When is Multicalibration Post-Processing Necessary?
por: Hansen, Dutch, et al.
Publicado: (2024)

Exploiting LLMs for Automatic Hypothesis Assessment via a Logit-Based Calibrated Prior
por: Gong, Yue, et al.
Publicado: (2025)

Position: Uncertainty Quantification Needs Reassessment for Large-language Model Agents
por: Kirchhof, Michael, et al.
Publicado: (2025)

Understanding the Emergence of Seemingly Useless Features in Next-Token Predictors
por: Rofin, Mark, et al.
Publicado: (2026)

Batch Calibration: Rethinking Calibration for In-Context Learning and Prompt Engineering
por: Zhou, Han, et al.
Publicado: (2023)

Calibrating LLMs for Text-to-SQL Parsing by Leveraging Sub-clause Frequencies
por: Liu, Terrance, et al.
Publicado: (2025)

Local Mechanisms of Compositional Generalization in Conditional Diffusion
por: Bradley, Arwen
Publicado: (2025)

Outliers and Calibration Sets have Diminishing Effect on Quantization of Modern LLMs
por: Paglieri, Davide, et al.
Publicado: (2024)

ROMER: Expert Replacement and Router Calibration for Robust MoE LLMs on Analog Compute-in-Memory Systems
por: Zhou, Wenyong, et al.
Publicado: (2026)

Low-Bit Quantization Favors Undertrained LLMs: Scaling Laws for Quantized LLMs with 100T Training Tokens
por: Ouyang, Xu, et al.
Publicado: (2024)

A Semantic-Sampling Framework for Evaluating Calibration in Open-Ended Question Answering
por: Wang, Zhanliang, et al.
Publicado: (2026)

Shielded Diffusion: Generating Novel and Diverse Images using Sparse Repellency
por: Kirchhof, Michael, et al.
Publicado: (2024)

LLMs and the Madness of Crowds
por: Bradley, William F.
Publicado: (2024)

LoCa: Logit Calibration for Knowledge Distillation
por: Yang, Runming, et al.
Publicado: (2024)

Geometry-Calibrated Conformal Abstention for Language Models
por: Xu, Rui, et al.
Publicado: (2026)

QA-Calibration of Language Model Confidence Scores
por: Manggala, Putra, et al.
Publicado: (2024)

On the Importance of a Multi-Scale Calibration for Quantization
por: Son, Seungwoo, et al.
Publicado: (2026)

Aligning LLMs with Human Uncertainty: A Beta-Bernoulli Calibrator for LLM Forecasting
por: Dai, Hui, et al.
Publicado: (2026)