The paper introduces Relative Bias to quantify LLM bias systematically.
problem Quantifying bias in LLMs is challenging due to ambiguity and rapid model emergence.
method Relative Bias framework using Embedding Transformation and LLM-as-a-Judge methodologies.
result The two scoring methods show strong alignment, providing a systematic approach.
Financial LLMs need explicit bias consideration to avoid invalid results.
problem Finance-specific biases inflate performance and contaminate backtests.
method Identified five recurring biases and proposed a Structural Validity Framework.
result Explicit bias consideration is necessary for valid deployment claims.
Fine-tunes LLMs to correct bias in predictions.
problem LLMs exhibit bias in predictions from data.
method Supervised fine-tuning with Low-Rank Adaptation (LoRA).
result Fine-tuning corrects bias in both controlled and real-world settings.
We detect lookahead bias in LLM forecasts using a novel statistical method.
problem Detecting lookahead bias in LLM-generated economic forecasts.
method Developed a statistical procedure using date-only recall queries and estimated Lookahead Propensity (LAP).
result LLM forecasts are contaminated with lookahead bias, as indicated by a positive interaction between LAP and the forecast in accuracy regressions.
Look-Ahead-Bench evaluates financial LLMs for lookahead bias, revealing significant differences in model performance.
problem Measuring and mitigating lookahead bias in financial LLMs.
method Standardized benchmark evaluating model behavior in practical financial scenarios, analyzing performance decay across market regimes.
result Standard LLMs exhibit significant lookahead bias, while Pitinf models show improved generalization and reasoning abilities.
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.
MultiwayPAM clusters LLM-as-a-Judge scores to reveal evaluator bias.
problem Revealing bias in LLM evaluations of text quality.
method Tensor clustering with MultiwayPAM method.
result Medoids reveal cluster membership of questions, answerers, and evaluators.
Study evaluates if LLMs have company-specific biases in financial sentiment analysis.
problem Evaluating if large language models exhibit company-specific biases in financial sentiment analysis.
method Comparing sentiment scores with and without company names, constructing economic models, and empirical analysis.
result LLMs show company-specific biases in sentiment analysis, impacting investor behavior and stock prices.
LLMs show biases in investment analysis, leading to unreliable recommendations.
problem LLMs face conflicts between pre-trained knowledge and real-time market data, leading to biases in investment analysis.
method Experimental framework to investigate emergent behaviors in LLMs, analyzing sector, size, and momentum biases.
result Distinct, model-specific biases observed, including a tendency to prefer technology stocks, large-cap stocks, and contrarian strategies.
AI bias arises from human-defined goals, not algorithmic flaws.
problem AI bias due to human-defined goals in LLMs.
method Purpose-conditioned cognition and revealing downstream use of LLM outputs.
result AI bias can be reduced by purpose-aware prompting but not fully by regularization.
New RLHF approach mitigates bias in aligning LLMs with human preferences.
problem Algorithmic bias in RLHF leading to preference collapse.
method Preference Matching (PM) RLHF, using PM regularizer and conditional variant.
result 29% to 41% improvement in alignment with human preferences.
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.
Anonymizing company names in financial news improves trading performance, contrary to initial expectations.
problem Look-ahead and distraction biases in sentiment analysis of financial news.
method Investigated trading strategies based on original and anonymized headlines, comparing performance.
result Anonymized headlines outperform original in-sample, suggesting distraction effect is stronger.
CausalGame benchmarks LLM agents' causal thinking in games.
problem Evaluating causal thinking in AI Scientists with LLMs.
method Interactive games with 14 scenarios incorporating selection bias, measurement error, and hidden confounders.
result None of the 30 LLM agents demonstrated reliable causal thinking, with the best model achieving only 68.0% survival.
Fairness in LLMs is impossible due to inherent technical challenges.
problem Ensuring fairness in large language models (LLMs) with rigorous definitions.
method Analysis of various technical fairness frameworks.
result No feasible technical fairness frameworks for LLMs due to large amounts of unstructured data and many potential combinations.
PAC-MCTS addresses biased search in LLM-guided planning by dynamically pruning.
problem Systematic biases in LLMs lead to inefficient and unsafe search in deep planning tasks.
method Formulates node expansion as BAI under bounded bias, derives sample complexity bounds, and proposes PAC-MCTS for dynamic confidence bounds.
result PAC-MCTS improves robustness and efficiency by up to 78% fewer API evaluations and 3x higher sample efficiency.
Study detects and explains positional bias in financial LLMs.
problem Positional bias in financial decision-making using LLMs.
method Unified framework and benchmark for detecting and quantifying bias in Qwen2.5 models.
result Positional bias is pervasive, scale-sensitive, and resurfaces under nuanced prompt designs.
Measures faithfulness of LLM explanations to reveal hidden biases and misleading claims.
problem LLM explanations can misrepresent the model's reasoning process, leading to over-trust and misuse.
method Defines faithfulness in terms of concept influence and uses counterfactuals and Bayesian models to estimate it.
result Can quantify and discover interpretable patterns of unfaithfulness in LLM explanations.
New method for evaluating LLMs reduces bias in open-ended evaluations.
problem Bias in Elo-based ratings of LLMs due to data redundancies.
method Proposes evaluation as a 3-player game and introduces novel solution concepts.
result Novel method leads to more robust and intuitive ratings.
A new method uses LLMs to discover causal pathways that affect fairness in machine learning.
problem Discovering fairness-relevant causal pathways in the presence of noise and confounding.
method Hybrid LLM-guided causal discovery framework combining active learning and dynamic scoring.
result LLM-guided methods, including the proposed active, dynamically scored variant, outperform baselines in recovering fairness-relevant structure under noisy conditions.
A simple method reduces bias in LLM auto-evaluators by controlling output length.
problem Bias in LLM auto-annotators, particularly length bias.
method Simple regression analysis to control for length difference and other mediators.
result Improved robustness and increased correlation with human preferences.
CARE improves LLM aggregation by accounting for shared confounders.
problem LLM judges' correlated errors due to shared latent confounders.
method CARE explicitly models judges' scores as true quality and confounders.
result CARE reduces aggregation error by up to 26.8% across various benchmarks.
Study proposes a multi-agent framework to mitigate bias in sentiment analysis.
problem Bias in sentiment analysis models.
method Integrates multiple LLMs, incorporates dialogue sessions, and uses probabilistic prediction.
result KCS+IBC reduces entropy and increases variance, suggesting improved balance between aggregation and diversity.
Benchmark assesses LLMs' causal inference skills, revealing significant limitations.
problem Lack of rigorous evaluation of LLMs' causal inference capabilities.
method CausalPitfalls benchmark with structured challenges and grading rubrics.
result Significant limitations in current LLMs' statistical causal inference.
A blindfolded LLM trading framework validates market signals without ticker memorization.
problem Ensuring LLMs trade based on genuine market understanding, not memorized data.
method Anonymize tickers and company names, verify signals through reasoning embeddings, and use PPO-DSR policy.
result Achieved Sharpe ratio of 1.40 +/- 0.22 across 20 seeds, robust in volatile markets.
Optimizes LLM prompts using logged user feedback.
problem Naive approaches to optimizing LLM prompts suffer from high variance and bias.
method Kernel-based off-policy gradient method leveraging sentence similarity.
result Substantially reduces variance and suppresses bias in optimizing prompts.
Paper proposes a data augmentation method for LLM-generated data in market research.
problem Bias in LLM-generated data in market research.
method Statistical data augmentation approach integrating LLM-generated and real data.
result Statistically robust estimators with reduced bias and cost savings.
Telescope detects LLM generated text by measuring token repetition probability.
problem Distinguishing LLM generated text from human writing.
method Telescope Perplexity, evaluating token repetition probability.
result Telescope Perplexity enables effective zero-shot LLM detection.
New SAE algorithm proves feature recovery for LLMs with theoretical guarantees.
problem Achieving interpretable features in large language models (LLMs).
method Proposed a statistical framework and bias adaptation technique for sparse autoencoders (SAEs).
result Proved correct recovery of all monosemantic features under specific data sampling.
R-AutoEval+ improves model evaluation efficiency and reliability using adaptive synthetic data.
problem Accurate model selection from AI candidates using real-world data is costly and impractical at scale.
method R-AutoEval+ uses adaptive prediction-powered inference to correct bias in autoevaluators while maintaining or improving sample efficiency.
result R-AutoEval+ provides finite-sample reliability guarantees and enhanced sample efficiency compared to conventional methods.
LLM evaluation suffers from systematic biases and lacks reliable positive judgments.
problem LLM evaluation suffers from systematic biases and lacks reliable positive judgments.
method Formulate LLM evaluation as a positive-unlabelled learning problem and propose a geometric auditing framework based on Partial Optimal Transport.
result Improved alignment with human preferences, increased robustness to presentation biases, and interpretable confidence estimates.
Inverse depth scaling found in LLMs due to similar layers averaging error.
problem Understanding how depth affects loss in large language models.
method Analysis of LLMs and toy residual networks.
result Loss scales inversely proportional to depth in LLMs.
Meta-Router optimizes LLM selection using gold-standard and preference-based data.
problem Training a high-quality LLM router with combined data sources is challenging due to bias and scarcity.
method Developed an integrative causal router training framework to correct bias and improve routing accuracy.
result Our approach delivers more accurate routing and improves the trade-off between cost and quality.
New benchmark uncovers hidden biases in LLMs that refuse to answer certain queries.
problem Evaluating fairness in LLMs, especially in sensitive applications.
method Silenced Bias Benchmark (SBB) using activation steering to reduce model refusals during QA.
result Exposes hidden unfair preferences in LLMs' latent space, distinguishing direct responses from underlying fairness issues.
FinReflectKG - EvalBench benchmarks financial KG extraction from SEC 10-K filings.
problem Lack of universal benchmark and evaluation framework for financial KG construction.
method Agentic and holistic evaluation principles, deterministic commit-then-justify judging protocol, binary and ordinal evaluations.
result Reflection-based extraction outperforms single-pass extraction in comprehensiveness, precision, and relevance.
AURA: Adaptive Uncertainty-aware Refinement for LLM-as-a-Judge Auditing
problem Auditing LLM-as-a-Judge decisions
method Adaptive uncertainty-aware refinement
result Human-consistent signal learning and evidence propagation
Framework assesses autograders' reliability and biases.
problem Mixed reliability and biases in autograders for LLM evaluation.
method Bayesian GLMs to model evaluation outcomes.
result Explicit quantification of scoring differences and biases.
Dynamic Vocabulary Pruning stabilizes LLM training by removing low-probability tokens.
problem Training Large Language Models (LLMs) with Reinforcement Learning (RL) causes numerical divergence between inference and training.
method Dynamic Vocabulary Pruning (DVP) constrains the RL objective to a safe vocabulary that excludes low-probability tokens.
result DVP stabilizes training by reducing systematic bias introduced by the extreme tail of the token distribution.
Language model benchmarks often misrepresent true understanding, revealing vulnerabilities in evaluation methods.
problem Language model benchmarks fail to accurately reflect true language understanding and adaptability.
method Systematic analysis of NLP evaluation frameworks, identifying vulnerabilities in static benchmarks, human evaluation protocols, and LLM-as-judge frameworks.
result Current evaluation methods are unreliable and need improvement to accurately assess LLM performance.
This paper uses LLMs for causal discovery with active learning and dynamic scoring to improve efficiency and fairness.
problem High computational demands and complexities of large-scale data in causal discovery.
method Metadata-based approach, BFS strategy, Active Learning, Dynamic Scoring Mechanism, LLM confidence scores.
result Significantly reduced number of queries and improved efficiency in causal graph construction.
Large language models struggle with causal relationships, leading to biases and hallucinations.
problem LLMs struggle with true causal relationships, leading to biases and hallucinations.
method Embed causality into LLMs training process at every stage.
result LLMs need to be trained to understand and apply causal knowledge, not just recite it.
Algorithm reduces audit costs by identifying best service configurations from biased textual evidence.
problem Designing service systems from textual evidence requires accurate selection despite biased automated scoring.
method Developed PP-LUCB algorithm combining LLM scores and selective audits to minimize costs.
result Correctly identified the best model in 40/40 trials with 90% cost reduction.
This paper tackles interpretability of LLMs in finance.
problem Complexity and lack of transparency of LLMs in finance.
method Mechanistic interpretability to understand LLM behavior.
result Demonstrates practical relevance of mechanistic interpretability in financial use cases.
We rebias estimates to improve interval calibration and prediction accuracy.
problem Constructing accurate intervals for noisy and biased estimates.
method Empirical Bayes rebiasing strategy that learns bias distribution from data.
result Substantial precision gains in prediction-powered inference.
Unified framework for certifying LLM reliability without extra supervision.
problem Improving reliability of large language models without additional supervision.
method Unified framework using majority voting and Martingale Majority Certificate (MMC).
result Certifiable inference in LLMs with statistical guarantees and adaptive stopping rules.
This paper uses LLMs to improve equity stock ratings by ingesting diverse financial and news data.
problem Challenges in traditional stock rating methods, including data overload, inconsistencies, and delayed reactions.
method Application of LLMs to generate multi-horizon stock ratings using various datasets.
result LLMs enhance the accuracy and consistency of stock ratings, outperforming traditional methods in forward returns.
Study shows training duration impacts model merging quality, suggesting joint selection of duration and method.
problem Impact of expert training duration on model merging quality for large language models (LLMs).
method Systematically fine-tuned experts on five domains across three model sizes, evaluating five merging methods at each duration.
result Training duration affects merging quality, with simple averaging degrading sharply and sparsification-based methods performing well past the validation optimum.
LLMs struggle to generate random numbers from statistical distributions, leading to biased results in applications.
problem LLMs' inability to generate random numbers accurately from specified distributions.
method Dual-protocol design: Batch Generation and Independent Requests, benchmarking 11 models across 15 distributions.
result Sampling fidelity degrades with distributional complexity and horizon, leading to systematic biases in downstream applications.