THERMOMETER calibrates LLMs efficiently for diverse tasks.
problem Calibrating large language models is challenging due to computational and versatility issues.
method THERMOMETER learns an auxiliary model for calibrating a LLM using data from multiple tasks.
result THERMOMETER produces better-calibrated responses for new tasks.
LLMs show surprising confidence in their answers, beyond just tokens.
problem LLMs lack meaningful confidence estimates for their responses.
method Semantic calibration test based on local loss optimality and equivalence classes.
result Base LLMs are semantically calibrated across tasks, contrary to expectations.
Corrects bias in LLM-as-a-judge evaluations using adaptive calibration.
problem Bias in LLM evaluations due to imperfect sensitivity and specificity.
method Plug-in framework with confidence intervals accounting for test and calibration dataset uncertainties.
result LML-based evaluation yields more reliable estimates than human-only evaluation.
The paper addresses poor calibration in fine-tuned LLMs after preference alignment.
problem Poor calibration in fine-tuned Large Language Models (LLMs) after preference alignment.
method Proposes a calibration-aware fine-tuning approach to restore calibration without compromising model performance.
result Demonstrates the effectiveness of the proposed methods through extensive experiments.
SC unifies ICL calibration methods and improves LLM performance.
problem Systematic biases in LLM predictions leading to unstable performance.
method Supervised Calibration (SC) learns optimal affine transformations in logit space.
result SC delivers state-of-the-art performance across multiple datasets.
A new framework evaluates LLM calibration in open-ended QA.
problem Evaluating LLM calibration in open-ended QA settings.
method Sem-ECE framework: sampling answers, grouping by semantic classes, and using frequencies as confidence.
result Sem-ECE estimators are unbiased and Sem2 achieves smaller calibration error. New methods correct bias in LLM-as-a-Judge evaluations, but reliability depends on judge quality and model calibration.
problem Systematic bias in LLM-as-a-Judge evaluations using naive estimators.
method Analytical results, simulations, and real-data case study to diagnose reliability of corrected estimates.
result Corrected estimates, especially shared-calibration comparisons, can be unreliable under certain conditions.
The study examines label smoothing to improve confidence calibration in fine-tuned LLMs.
problem Improving confidence calibration in fine-tuned large language models (LLMs) after instruction tuning.
method Examine various open-sourced LLMs, label smoothing, and custom kernel design.
result Label smoothing is effective in maintaining confidence calibration but faces challenges in large vocabulary LLMs.
Study shows multilingual LLM calibration effects improve model confidence but not accuracy.
problem Improving multilingual language model calibration in low-resource settings.
method Analysis of two multilingual benchmarks using instruction-tuning and label smoothing.
result Model confidence increases in low-resource languages after instruction-tuning but accuracy improvements are marginal.
CJE calibrates cheap LLM judges against an oracle, achieving high accuracy at a fraction of the cost.
problem Inexpensive LLM judges can produce biased rankings, leading to unreliable outcomes.
method CJE uses a small oracle to calibrate cheap scores, then evaluates at scale with valid uncertainty.
result CJE achieves 99% pairwise ranking accuracy at 14x lower cost compared to a 16x oracle/judge cost ratio.
Develops a power-calibrated framework for LLM watermarking, optimizing tradeoffs between detectability and distortion.
problem The trade-off between detectability and semantic distortion in logit-based watermarking.
method Power-calibrated statistical framework for watermark hyperparameters, establishing explicit relationships.
result Derives practical parameter selection procedures achieving optimal tradeoffs under constraints.
The paper introduces multicalibration to improve confidence scores in LLMs.
problem Improving the reliability and interpretability of confidence scores for LLMs.
method Forming groupings of prompt/completion pairs correlated with correctness, using clustering and self-annotation. Developing multicalibration algorithms to reduce overfitting.
result Our techniques yield confidence scores that significantly improve calibration and accuracy compared to existing methods.
This work shows fine-tuning improves LLM uncertainty estimates efficiently.
problem Ensuring high-stakes predictions from large language models.
method Fine-tuning on a small dataset of correct and incorrect answers.
result Fine-tuned LLMs produce good generalization and tractable uncertainty estimates.
Calibrated PRMs improve inference efficiency for LLMs by dynamically adjusting compute budgets.
problem Poor calibration of PRMs leads to overestimation of success probabilities in partial reasoning steps.
method Quantile regression for calibration, instance-adaptive scaling (IAS) framework.
result Calibrated PRMs reduce inference costs while maintaining accuracy, especially on confident problems.
Cost-effective method detects language model hallucinations.
problem Detecting unreliable outputs from LLMs.
method Pipeline including confidence score, input attributes calibration, and thresholding.
result Multi-scoring framework outperforms individual methods and reduces computational cost.
G-Sim uses LLMs to build reliable simulators for complex systems.
problem Building robust simulators for critical domains like healthcare and logistics is challenging.
method Hybrid framework combining LLM-driven structural design and empirical calibration.
result G-Sim produces reliable, causally-informed simulators that handle non-differentiable and stochastic simulators.
MACI improves LLM factuality inference with higher retention and lower time cost.
problem Ensuring factuality in LLM responses for high-stakes domains.
method Reformulated conformal inference in a multiplicative filtering setting, leveraging ensembles for more accurate factuality scores and group-conditional calibration.
result MACI achieves higher retention and lower time cost compared to baselines, preserving validity through group-conditional calibration.
We develop a method to estimate the time to unsafe responses in LLMs.
problem Estimating the time to unsafe responses in large language models is challenging due to the rarity of unsafe outputs.
method We frame the problem as survival analysis and propose a calibration technique for constructing a lower predictive bound (LPB).
result Our method provides rigorous coverage guarantees and improves sample efficiency.
Framework controls uncertainty in LLMs without labels or probabilities.
problem Managing uncertainty in black-box LLMs without token-level probability or true labels.
method Integrates generative models, UCP, and conformal alignment to control uncertainty.
result Achieves close-to-nominal coverage and tighter thresholds than split UCP.
ABC improves uncertainty quantification in LLMs for clinical diagnostics.
problem Overconfident and poorly calibrated estimates of LLMs in clinical domains.
method Approximate Bayesian Computation (ABC) for likelihood-free inference.
result Improves accuracy by up to 46.9%, reduces Brier scores by 74.4%, and enhances calibration.
New method calibrates LLMs for safety-critical tasks with scalable Bayesian inference.
problem Overconfidence in LLMs after fine-tuning for specific tasks.
method Orthogonalized Low-Rank Adapters (PoLAR) with variational Bayesian inference.
result Scalable and well-calibrated uncertainty estimation for LLMs.
LLMs overestimate stock returns and are less accurate at predicting extreme outcomes.
problem Behavioral biases in LLMs' stock return forecasts.
method Comparison of LLM forecasts with crowd-sourced estimates and historical data.
result LLMs overestimate stock returns and are less accurate at predicting extreme outcomes.
Unified framework to bridge human and LLM judgments.
problem Systematic discrepancies between human and LLM evaluations.
method Latent human preference score and linear transformations of covariates.
result Higher agreement with human ratings and exposure of systematic gaps.
Simple online monitor detects unsafe LLM outputs.
problem LLMs generate unsafe outputs despite training.
method Thresholding external verifier signal to decide alarms.
result Simple design competitive with advanced methods.
Novel approach trains LLMs for inductive reasoning using probabilistic programs.
problem Training LLMs for inductive reasoning with sparse, ambiguous data.
method Program-based Posterior Training (PPT) using probabilistic inference.
result Significant improvement in estimation accuracy and alignment with human judgments.
Variance-Calibrated Modulation (VCM) addresses the likelihood trap in LLMs by reshaping the probability distribution before truncation.
problem LLMs fall into the likelihood trap, leading to repetitive degeneration and vocabulary dullness.
method VCM reshapes the probability distribution before truncation through Contextual Searchlight and Adaptive Self-Debiasing.
result VCM mitigates the likelihood trap across open-ended generation, factual QA, and mathematical reasoning.
ORCA calibrates LLMs for efficient, generalizable reasoning.
problem Miscalibration of large language models leading to inefficiencies.
method Online Reasoning Calibration (ORCA) using conformal prediction and test-time training.
result ORCA provides higher efficiency and generalization across different reasoning tasks.
Temperature scaling improves model uncertainty but not diversity in LLMs.
problem Improving the calibration and stochasticity of probabilistic models.
method Investigates theoretical properties of temperature scaling in classification and LLMs.
result Temperature scaling increases model uncertainty but not diversity in LLMs.
ValueBlindBench tests LLM-generated investment rationales for validity before returns are known.
problem Delayed-ground-truth evaluation of LLM-generated investment rationales.
method Agreement-gated stress testing protocol to validate LLM-judged rationales.
result ValueBlindBench prevents overclaims and identifies flawed financial constructs.
Statsformer validates and adapts LLM-derived semantic priors for improved supervised learning.
problem Unreliable semantic priors from LLMs can degrade supervised learning performance.
method Adapts LLM-derived feature scores into a family of learner-specific prior-injection mechanisms, calibrating their influence using out-of-fold validation.
result Improves prediction performance by adaptively downweighting unreliable LLM priors, ensuring a guardrailed statistical learning system.
Bayesian MoE framework improves LLMs' uncertainty detection.
problem Brittleness and overconfidence in deterministic routing of LLMs.
method Structured Bayesian routing in weight-space, logit-space, and selection-space.
result Significant improvements in routing stability, calibration, and OoD detection.
Low-rank framework for task-specific LLM ranking from sparse comparisons.
problem Challenges in reliable task-specific ranking of LLMs under sparse, imbalanced comparisons.
method Low-rank modeling of task-by-model ability matrix, max-norm accurate estimator, task-wise top-K recovery guarantees, uncertainty quantification framework.
result Improves sample efficiency and produces tighter, better-calibrated ranking certificates.
New method for LLMs to learn reasoning by optimizing latent variables.
problem Teaching LLMs to generate logical justifications for answers.
method Formalized reasoning as latent variable model, derived FEM objective, designed sampling schemes.
result Prompt Posterior Sampling (PPS) outperforms other schemes in learning to reason.
New method corrects quantization errors in LLMs using low-rank matrices.
problem Correcting quantization errors in large language models.
method Introducing low-rank weight matrices to correct quantized activations in LLMs.
result Reduces accuracy gap with original model by more than 50% using low-rank matrices.
Develops a framework for quantifying agentic AI model risk using LLM-inferred Bayesian state filters.
problem Quantifying the risk of agentic AI systems due to uncertain beliefs and actions.
method Representing the system as a partially observed Markov decision process with latent states, Bayesian belief updates, control-dependent losses, and tail-risk functionals.
result Develops a rigorous framework for separating uncertainty quantification from risk measurement.
Bayesian approach quantifies uncertainty in LLM-based systems.
problem Uncertainty quantification in LLM-based systems, especially for high-stakes applications.
method Interpreting prompts as parameters in a Bayesian model, using MHLP for inference.
result Improvements in predictive accuracy and uncertainty quantification on various benchmarks.
Improved text-conditioned regression using LLMs and diffusion-based neural processes.
problem Major error cascades and computational inefficiency in LLMs for short sequences.
method Combining LLM predictive densities with a diffusion-based neural process.
result Better-calibrated predictions and locally consistent trajectories.
New method samples from LLM posterior for coherent, useful responses.
problem Hallucinations in large language models.
method Posterior sampling for conditional generation, with calibration.
result Achieves statistical guarantees with higher downstream utility.
A new framework evaluates LLMs by considering judge reliability.
problem Evaluating LLMs without ground truth labels can lead to biased results.
method Introduces judge-specific discrimination parameters and estimates model quality and judge reliability.
result Improves agreement with human preferences and produces calibrated uncertainty quantification.
Paper explores using LLMs for zero-shot reinforcement learning in continuous spaces.
problem Leveraging LLMs for continuous state spaces in reinforcement learning.
method Disentangled In-Context Learning (DICL) to handle multivariate data and control signal.
result DICL produces well-calibrated uncertainty estimates in reinforcement learning settings.
New approach treats coordination as an architectural layer to improve LLM-based multi-agent systems.
problem Coordination defects lead to high failure rates in LLM-based multi-agent systems.
method Treats coordination as a configurable architectural layer separable from agent logic and information access.
result Configurations leave distinguishable signatures, enabling architectural reasoning and Pareto frontiers.
Study improves LLMs for PPI analysis by addressing uncertainty.
problem Uncertainty in LLM predictions for PPIs.
method Fine-tuned LLaMA-3 and BioMedGPT models, LoRA ensembles, Bayesian LoRA for UQ.
result Competitive PPI identification performance across diverse disease contexts.
FinTradeBench benchmarks LLMs for financial reasoning combining company fundamentals and market signals.
problem Challenges in evaluating financial reasoning models for LLMs.
method Developed a benchmark integrating company fundamentals and trading signals, using a calibration-then-scaling framework.
result Clear performance gap between LLMs, retrieval improves reasoning over textual fundamentals but not trading signals.
PRCD-MAP learns to trust imperfect priors in causal discovery, improving accuracy and robustness.
problem Tackles the brittle trade-off between blind trust and rejection of external priors in causal discovery.
method Proposes PRCD-MAP, a soft prior-consumption layer that assigns per-edge trust to imperfect priors and modulates regularization in a MAP objective.
result Enjoys a population-level safety guarantee and outperforms existing methods on real-world causal discovery tasks.
PolySwarm uses a swarm of LLMs to predict and arbitrage prediction markets.
problem Real-time prediction market trading and latency arbitrage inefficiencies.
method PolySwarm employs a swarm of 50 diverse LLMs, Bayesian combination, and risk-controlled execution.
result Swarm aggregation outperforms single-model baselines in prediction tasks.
Bayesian principles improve agentic AI decision-making.
problem Decision-making under uncertainty in agentic AI systems.
method Bayesian decision theory applied to the orchestration layer of agentic AI.
result Bayesian principles enhance agentic AI's ability to make decisions under uncertainty.
Thinking LLMs struggle with stock prediction, especially as data complexity increases.
problem Evaluating the performance of 'thinking' LLMs in stock prediction, especially under varying levels of cross-sectional complexity.
method Rolling 48m/1m walk-forward evaluation, comparing direct LLMs, TLLMs, and classical learners on cross-sectional ranking loss, MSE, and backtests with transaction costs.
result TLLMs' ranking quality deteriorates as cross-sectional complexity grows, while direct LLMs remain stable.
LLMs cause inconsistent financial outputs, smaller models are more reliable.
problem Inconsistent outputs from LLMs undermine auditability and trust in financial workflows.
method Finance-calibrated deterministic test harness, task-specific invariant checking, model classification, and cross-provider validation.
result Smaller models (Granite-3-8B, Qwen2.5-7B) achieve 100% output consistency, while larger models like GPT-OSS-120B have high drift.