Cover Image

Sampling More, Getting Less: Calibration is the Diversity Bottleneck in LLMs

Fuente: arXiv

Saved in:

Bibliographic Details
Main Authors:	Banayeeanzade, Amin, Yang, Qingchuan, Tarsadiya, Dhruv, Bahrani, Fatemeh, Blas, Leonardo, Samuel, Alfy, Jia, Robin, Razaviyayn, Meisam, Karimireddy, Sai Praneeth
Format:	Preprint
Published:	2026
Subjects:	Computation and Language
Online Access:	Acceder al recurso
Tags:	Add Tag No Tags, Be the first to tag this record!

Similar Items

Psychological Steering in LLMs: An Evaluation of Effectiveness and Trustworthiness
by: Banayeeanzade, Amin, et al.
Published: (2025)

EPSVec: Efficient and Private Synthetic Data Generation via Dataset Vectors
by: Banayeeanzade, Amin, et al.
Published: (2026)

Sparks of Rationality: Do Reasoning LLMs Align with Human Judgment and Choice?
by: Tak, Ala N., et al.
Published: (2026)

f-INE: A Hypothesis Testing Framework for Estimating Influence under Training Randomness
by: Panda, Subhodip, et al.
Published: (2025)

GABRIL: Gaze-Based Regularization for Mitigating Causal Confusion in Imitation Learning
by: Banayeeanzade, Amin, et al.
Published: (2025)

ContextLeak: Auditing Leakage in Private In-Context Learning Methods
by: Choi, Jacob, et al.
Published: (2025)

Less is More: Convergence Benefits of Fewer Data Weight Updates over Longer Horizon
by: Das, Rudrajit, et al.
Published: (2026)

Output Perturbation for Differentially Private Convex Optimization: Faster and More General
by: Lowy, Andrew, et al.
Published: (2021)

Beyond URLs: Metadata Diversity and Position for Efficient LLM Pretraining
by: Fan, Dongyang, et al.
Published: (2025)

AutoFocus-IL: VLM-based Saliency Maps for Data-Efficient Visual Imitation Learning without Extra Human Annotations
by: Gong, Litian, et al.
Published: (2025)

Optimization with Access to Auxiliary Information
by: Chayti, El Mahdi, et al.
Published: (2022)

Private Federated Learning Without a Trusted Server: Optimal Algorithms for Convex Losses
by: Lowy, Andrew, et al.
Published: (2021)

Private Stochastic Optimization With Large Worst-Case Lipschitz Parameter
by: Lowy, Andrew, et al.
Published: (2022)

On the Limits of Momentum in Decentralized and Federated Optimization
by: Zaccone, Riccardo, et al.
Published: (2025)

Ghosted Layers: Unconstrained Activation Alignment for Recovering Layer-Pruned LLMs
by: Yun, Vincent-Daniel, et al.
Published: (2026)

Sampling and Loss Weights in Multi-Domain Training
by: Salmani, Mahdi, et al.
Published: (2025)

Collaborative Heterogeneous Causal Inference Beyond Meta-analysis
by: Guo, Tianyu, et al.
Published: (2024)

Defection-Free Collaboration between Competitors in a Learning System
by: Werner, Mariel, et al.
Published: (2024)

Do Data Valuations Make Good Data Prices?
by: Fan, Dongyang, et al.
Published: (2025)

Uncertainty as Feature Gaps: Epistemic Uncertainty Quantification of LLMs in Contextual Question-Answering
by: Bakman, Yavuz, et al.
Published: (2025)

Robust Multi-Agent LLMs under Byzantine Faults
by: Lee, Haejoon, et al.
Published: (2026)

LIA: Privacy-Preserving Data Quality Evaluation in Federated Learning Using a Lazy Influence Approximation
by: Rokvic, Ljubomir, et al.
Published: (2022)

Communication-Efficient Heterogeneous Federated Learning with Generalized Heavy-Ball Momentum
by: Zaccone, Riccardo, et al.
Published: (2023)

A Differentially Private Kaplan-Meier Estimator for Privacy-Preserving Survival Analysis
by: Veeraragavan, Narasimha Raghavan, et al.
Published: (2024)

Policy Gradient Converges to the Globally Optimal Policy for Nearly Linear-Quadratic Regulators
by: Han, Yinbin, et al.
Published: (2023)

f-FERM: A Scalable Framework for Robust Fair Empirical Risk Minimization
by: Baharlouei, Sina, et al.
Published: (2023)

Neural Network-Based Score Estimation in Diffusion Models: Optimization and Generalization
by: Han, Yinbin, et al.
Published: (2024)

Stochastic Control for Fine-tuning Diffusion Models: Optimality, Regularity, and Convergence
by: Han, Yinbin, et al.
Published: (2024)

Differentially Private Next-Token Prediction of Large Language Models
by: Flemings, James, et al.
Published: (2024)

On the Inherent Privacy of Zeroth Order Projected Gradient Descent
by: Gupta, Devansh, et al.
Published: (2025)

Adaptively Private Next-Token Prediction of Large Language Models
by: Flemings, James, et al.
Published: (2024)

DAVED: Data Acquisition via Experimental Design for Data Markets
by: Lu, Charles, et al.
Published: (2024)

Sample, Align, Synthesize: Graph-Based Response Synthesis with ConGrs
by: Ghosh, Sayan, et al.
Published: (2025)

Hair-Trigger Alignment: Black-Box Evaluation Cannot Guarantee Post-Update Alignment
by: Bakman, Yavuz, et al.
Published: (2026)

Data Dashboards: Get More with Less
by: Kerry Nenn
Published: (2025)

Differentially Private In-context Learning via Sampling Few-shot Mixed with Zero-shot Outputs
by: Flemings, James, et al.
Published: (2025)

Tradeoffs between convergence rate and noise amplification for momentum-based accelerated optimization algorithms
by: Mohammadi, Hesameddin, et al.
Published: (2022)

Hybrid Learners Do Not Forget: A Brain-Inspired Neuro-Symbolic Approach to Continual Learning
by: Banayeeanzade, Amin, et al.
Published: (2025)

Entropy-driven Fair and Effective Federated Learning
by: Wang, Lin, et al.
Published: (2023)

VoxGuard: Evaluating User and Attribute Privacy in Speech via Membership Inference Attacks
by: Tsaprazlis, Efthymios, et al.
Published: (2025)