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.
This paper integrates LLMs into SCD to improve causal inference accuracy.
problem Challenges in acquiring domain expert knowledge for causal models.
method Statistical causal prompting (SCP) for LLMs and prior knowledge augmentation for SCD.
result LLM-KBCI and SCD augmented with LLM-KBCI approach ground truths more closely.
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.
Understanding optimal prompts for binary sequence predictors is challenging.
problem Finding good prompts for binary sequence predictors is difficult.
method Viewing prompting as finding the best conditioning sequence on a near-optimal sequence predictor, using empirical and statistical analysis.
result Optimal prompts can be better understood given the pretraining distribution, which is not usually available.
This work analyzes CoT prompting methods from a statistical estimation perspective.
problem Improving the effectiveness of LLMs in solving multi-step reasoning problems.
method Introducing a multi-step latent variable model to characterize CoT prompting from a statistical estimation viewpoint.
result The CoT estimator is equivalent to a Bayesian estimator when the pretraining dataset is large.
Single-head transformers with a single self-attention layer can approximate any sequence-to-sequence function and are efficient under certain conditions.
problem Statistical and computational limits of prompt tuning for transformer-based models.
method Investigation of single-head transformers with a single self-attention layer, proving universality and efficiency under SETH.
result Existence of almost-linear time prompt tuning inference algorithms under certain conditions.
AI uses language models to find instrumental variables quickly.
problem Finding valid instrumental variables is a challenging and heuristic process.
method Uses large language models to search for new instrumental variables through narratives and counterfactual reasoning.
result Demonstrates the effectiveness of multi-step and role-playing prompting strategies for LLMs.
MIP framework improves urban flow prediction by adapting to distribution shifts.
problem Distribution shifts in urban flow data make prediction models unreliable.
method Memory-enhanced Invariant Prompt learning with learnable memory bank.
result MIP ensures robust predictions by focusing on invariant features.
Efficiently evaluate generative models at the prompt level using tensor factorization.
problem Fine-grained evaluations of generative models are costly and often misaligned with human judgment.
method Tensor factorization model that merges cheap autorater data with a small set of human gold-standard labels.
result The method provides accurate and tight confidence intervals for model performance.
Unified framework for studying softmax attention under large prompts.
problem Challenges in theoretical analysis of softmax attention.
method Measure-based framework for finite and infinite prompts.
result Softmax attention converges to linear attention in the large-prompt regime.
The paper compares and optimizes estimators for treatment effects with observed confounders and mediators.
problem Estimating treatment effects with observed confounders and mediators.
method Investigates the linear Gaussian causal model, compares and optimizes estimators, and combines datasets.
result An optimal estimator outperforms the backdoor and frontdoor estimators by an unbounded constant factor.
LLMs detect market patterns through causal reasoning, not just temporal association.
problem Detecting structural market patterns in financial data.
method Obfuscation testing using the WHO-WHOM-WHAT framework.
result LLMs achieve 71.5% detection rate of market patterns without temporal context.
Study quantifies how LLMs capture higher-order statistical structure using cumulant expansion.
problem Understanding how LLMs internalize statistical structure during next-token prediction.
method Cumulant-expansion framework treating softmax entropy as perturbation around center distribution.
result Cumulants reveal distinct signatures for mathematical vs. general text prompts, quantifying feature-learning dynamics.
This work frames reward modelling from preferences as a causal problem.
problem Reward modelling from preference data for AI alignment.
method Causal inference approach to identify challenges and assumptions.
result Causally-inspired approaches improve model robustness.
A new model designs molecules with desired properties.
problem Finding molecules with optimal chemical or biological properties.
method A latent prompt Transformer model with three components: latent vector, molecule generation, and property prediction.
result The model achieves state-of-the-art performance on molecule design tasks.
LLM4Causal democratizes causal reasoning via fine-tuned LLMs.
problem Limited capability of LLMs in causal inference and interpretation.
method Fine-tuning an open-source LLM for causal tasks, proposing datasets for instruction tuning.
result LLM4Causal delivers end-to-end solutions for causal problems and interprets results easily.
The paper studies estimation rates for MoE models with a new prompt.
problem Estimating parameters in a softmax-contaminated MoE model.
method Analytic notion of distinguishability, minimax optimal estimation rates.
result Estimation rates are minimax optimal under distinguishability, but slower otherwise.
LLMs fail to match statistical ground truth despite stable run-to-run performance.
problem LLMs lack validation against statistical ground truth in automated scientific workflows.
method Introduced a behavioral evaluation framework for LLMs, separating four decision-making dimensions.
result LLMs can exhibit near-perfect stability but diverge from statistical ground truth.
Emergent misalignment is influenced by training dynamics, model priors, and data.
problem Emergent misalignment in models
method Exploring training dynamics, model priors, and data
result Activation deltas before and after narrow fine-tuning correlate with their similarities when measured with the last prompt-token activations.
Reinforce-Ada improves RL for language models by adaptively sampling difficult prompts.
problem Signal loss in RL for large language models due to undersampling.
method Adaptive sampling based on a non-linear RL objective, optimizing prompt difficulty weighting.
result Reinforce-Ada significantly outperforms uniform sampling, recovering lost signals and accelerating convergence.
Research aims to bridge statistical learning to causal models in AI.
problem Challenges in machine learning and AI related to causality.
method Transition from statistical learning to causal models.
result Progress in AI may require advances in causal modeling.
Casper uses causal graph neural networks to improve spatiotemporal time series imputation.
problem Imputing missing values in spatiotemporal time series with confounders and non-causal correlations.
method Casper introduces a novel Prompt Based Decoder (PBD) and Spatiotemporal Causal Attention (SCA) to block confounders and discover causal relationships.
result Casper outperforms baselines and effectively discovers causal relationships in spatiotemporal time series imputation.
New framework improves interpretability of trainable prompts.
problem Improving task-specific LLM performance with soft prompts remains a black-box method.
method Developed a theoretical framework for evaluating interpretability of trainable prompts, inspired new objective functions.
result Found a fundamental trade-off between interpretability and task performance in trainable prompts.
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.
Study on how attention in prompt-tuning affects large language models.
problem Limited theoretical understanding of prompt-tuning and attention in LLMs.
method Exploration of prompt-tuning for one-layer attention architectures, contextual mixture-models, and self-contained prompt-attention model.
result Softmax-prompt-attention is more expressive than self-attention and linear-prompt-attention under contextual data model.
AI needs causal inference to avoid being just a correlation machine.
problem AI's inability to distinguish correlation from causation.
method Develops a unified framework connecting various causal statistical estimators and proves a Statistical Necessity Theorem for causal generalization.
result AI systems without causal grounding are brittle and biased, highlighting the need for causal statistics.
Kernel smoothing improves LLM reasoning efficiency.
problem Efficiently estimate value functions with limited samples for reinforcement learning.
method Kernelized advantage estimation using classical nonparametric statistics.
result Improved policy optimization with accurate value and gradient estimation.
InfoPrompt improves soft prompt tuning by maximizing mutual information, leading to better performance.
problem High sensitivity of prompt tuning to initial conditions and insufficient task-relevant information.
method Develops an information-theoretic framework to maximize mutual information between prompts and model parameters, using novel loss functions.
result InfoPrompt accelerates convergence and outperforms traditional methods.
New method optimizes language model performance for test-time strategies.
problem Mismatch between training objectives and test-time deployment of large language models.
method Tail-Extrapolated estimators to approximate best-of-N performance from limited training rollouts.
result Improved performance of best-of-N deployment across various models and datasets.
Automates zero-shot classification by scoring and weighting prompts.
problem Improving zero-shot accuracy through prompt ensembling.
method Automatic prompt scoring and weighting method.
result Method outperforms existing techniques on various benchmarks.
PDO optimizes LLM prompts without labels, improving performance.
problem Optimizing prompts for LLMs without access to labeled data.
method Pairwise preference feedback, dueling bandits, Thompson Sampling, mutation.
result PDO identifies stronger prompts than label-free methods.
CAPO optimizes LLM prompts more efficiently and cost-effectively.
problem Costly and inefficient automatic prompt optimization for LLMs.
method Integrates AutoML techniques for evolutionary optimization of instructions and few-shot examples.
result Significantly improves prompt optimization efficiency and accuracy.
Study interpolating estimators for causal learning from observational data.
problem Learning causal models from observational data in complex model classes.
method Investigate min-norm interpolators and ridge-regularized regressors in a linearly confounded model.
result Interpolators cannot be optimal for causal learning under the principle of independent causal mechanisms, requiring stronger regularization.
Quriosity analyzes curiosity-driven questions from diverse sources.
problem Understanding and analyzing human curiosity-driven questions.
method Collection of 13.5K questions from various sources, development of an iterative prompt improvement framework.
result 42% of questions are causal, revealing unique linguistic properties.
The paper analyzes identifiability in ODE systems with hidden confounders.
problem Identifiability of ODE systems with hidden confounders.
method Systematic analysis of identifiability in linear ODE systems with hidden confounders, considering both no causal relationships and causal dependencies.
result Comprehensive identifiability analysis of ODE systems with hidden confounders, including causal dependencies.
Study on forecasting methods and their causal implications.
problem Understanding the difference between statistical and causal risks in forecasting models.
method Introduce causal learning theory for forecasting, obtain uniform convergence bounds for VAR models.
result First theoretical guarantees for causal generalization in time-series forecasting.
Diffusion LLMs can efficiently generate harmful prompts for adversarial testing.
problem Generating harmful prompts for adversarial testing is resource-intensive and costly.
method Transformed adversarial prompt optimization into an efficient inference task using pretrained Diffusion LLMs.
result Only a few conditional samples are required to generate harmful prompts with high reward.
PromptEval estimates LLM performance across many prompts, improving reproducibility.
problem Limited prompt templates affect LLM benchmark reproducibility.
method Estimates performance distribution across many prompts using borrowed strength.
result PromptEval accurately estimates performance quantiles with practical budget.
Statistical tests for fairness in admissions data reveal hidden patterns.
problem Simpson's paradox in admissions data hides true gender bias.
method Introduces a new statistical test based on Pearl's instrumental-variable inequalities.
result Statistical tests for fairness coincide with causal notions for the Berkeley admissions case.
New method uses sufficient statistics to infer causal relationships from observational data.
problem Inferring causal relationships from observational data with hidden variables.
method Information Bottleneck method applied to find functional sufficient statistics.
result New causal rules not obtainable from standard methods, validated on simulated and real data.
M-FISHER detects and adapts to streaming data shifts with statistical validity and stability.
problem Detecting and adapting to distributional shifts in streaming data.
method Constructs an exponential martingale from non-conformity scores and applies Ville's inequality for detection. Fisher-preconditioned updates for adaptation.
result Establishes M-FISHER as a principled approach for robust, anytime-valid detection and geometrically stable adaptation.
The study explores how Transformers predict the next token in a sequence.
problem Understanding the mechanism behind Transformers' autoregressive learning ability.
method Exploring the approximation ability of Transformers for next-token prediction through specific instances and a causal kernel descent method.
result Transformer models can learn context-dependent functions f for next-token prediction based on past and current observations. TSCD is an algorithm for causal discovery using second-order statistics.
problem Causal discovery
method Tensor-based Second-order Causal Discovery (TSCD)
result Identifiable causal order and parameters from logarithmic number of interventions
Transformer models can approximate smooth functions with prompts, enhancing LLMs' dynamic capabilities.
problem Lack of theoretical framework for prompt engineering in transformer models.
method Formal framework demonstrating transformer models can approximate β-times differentiable functions with prompts. result Transformer models can approximate β-times differentiable functions with arbitrary precision using appropriately structured prompts. CRL uses causality to build interpretable AI models from complex data.
problem Interpreting deep neural networks' implicit representations.
method Causal representation learning (CRL) synthesizing latent variable models, causal graphical models, and nonparametric statistics.
result CRL can improve interpretability of generative AI models.
A framework for optimizing prompt selection in generative language models.
problem Efficiently selecting prompts for generative language models.
method Two-stage framework using simulation optimization to maximize a pre-defined score.
result Consistency of the sequential evaluation procedure in the proposed framework.
Transformers learn low-dimensional target functions efficiently in-context.
problem Efficiently learning nonlinear target functions in-context using transformers.
method Nonlinear MLP layer in transformers optimized by gradient descent, focusing on single-index target functions.
result Transformers can learn target functions with low-dimensional structures efficiently in-context.
This paper explains how to optimize prompts for model adaptation.
problem Understanding and optimizing prompt tuning for model adaptation.
method Bayesian view and meta-learning to explain prompt optimization.
result Optimal prompting can be studied formally as conditioning Bayesian predictors.