Sloth predicts LLM performance using latent skills across families.
problem Variations in benchmark performance due to differences in training configurations and data processing across model families.
method Sloth uses publicly available benchmark data and assumes LLM performance is driven by latent skills influenced by model size and training tokens. It exploits correlations across benchmarks to provide accurate predictions.
result Sloth predicts LLM performance accurately and offers insights into scaling behaviors for complex tasks.
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.
AutoScale improves LLM pre-training by adjusting data mixtures at different scales.
problem Data mixtures that work well at small scales may not perform as well at larger scales.
method AutoScale uses a two-stage approach: fitting a model to predict loss under different compositions and extrapolating optimal compositions to larger scales.
result AutoScale accelerates convergence and improves downstream performance.
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
LLMs learn peaked distributions slowly due to power-law losses.
problem Slow convergence of loss in training large language models.
method Systematic analysis of toy models and empirical evaluation of LLMs.
result Power-law time scaling with an exponent of 1/3 for learning peaked distributions.
Fine-tuned open-source LLMs match or exceed closed-source models in social science research.
problem Limited scalability and high costs of large LLMs in social science research.
method Fine-tuning open-source models for specific tasks, exploring training set size effects, proposing hybrid workflow.
result Small, fine-tuned open-source LLMs achieve equal or superior performance to commercial alternatives.
Efficiently allocate budgets for LLM-assisted virtual screening to reduce costs.
problem Reducing the cost of evaluating alternatives in large-scale screening tasks.
method Propose a top-m greedy evaluation mechanism and the EFG-m algorithm for efficient budget allocation. result Prove that EFG-m is both sample-optimal and consistent in large-scale virtual screening. This paper studies activation sparsity in large language models, finding key trends and implications.
problem Activation sparsity in large language models (LLMs) can be improved for efficiency and interpretability.
method Proposes PPL-p% sparsity, analyzes trends with training data, width-depth ratio, and parameter scale. result ReLU is more efficient for sparsity than SiLU, and deeper architectures can improve sparsity.
This paper uses LLMs and cycle consistency for better machine translation evaluation.
problem Evaluating translation quality and LLM capabilities without ground truth.
method Generate translation candidates, back-translate, and evaluate cycle consistency.
result Larger LLMs or more inference passes improve cycle consistency.
New method predicts and optimizes test-time scaling for LLMs.
problem Lack of principled guidance on scaling LLMs efficiently.
method Tail-guided search to predict and allocate compute.
result SLG Search achieves higher rewards with less compute.
Law explains how LLMs learn to predict next tokens.
problem Understanding how LLMs process input data internally.
method Introduced a precise law governing token embeddings in LLMs.
result Each layer equally contributes to next-token prediction accuracy.
Local SGD outperforms conventional methods in LLM training.
problem Training large language models on distributed devices.
method Local SGD algorithm applied to distributed training.
result Local SGD achieves competitive results in LLM training.
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.
New scaling laws optimize model size, training, and inference for better performance.
problem Trade-off between model size and inference cost in modern LLMs.
method Train-to-Test (T2) scaling laws that jointly optimize model size, training tokens, and inference samples. result Optimal pretraining decisions shift into overtraining regime, leading to stronger performance.
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.
This study reveals the critical role of scale vectors in large language models, improving optimization and expressivity.
problem Understanding and optimizing the scale vectors in large language models.
method Systematic study of scale vectors from expressivity, optimization, and architectural perspectives; theoretical and empirical analysis of weight decay; proposing and evaluating improvements.
result Scale vectors improve optimization through a self-amplifying preconditioning effect and are beneficial for expressivity in certain architectures.
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.
LLMs generate answers under incomplete context, and their uncertainty should scale with missing information.
problem Evaluating the quality of LLM answers under incomplete context.
method A controlled framework with varying context availability, and two uncertainty measures (sampling-based confidence and response entropy) evaluated on SQuAD.
result Response entropy increases with context removal and explains more variance in accuracy than confidence, suggesting it is a more responsive uncertainty measure.
Scales attention for long contexts in LLMs.
problem Development of attention mechanisms for long context inference.
method Scale-invariant total attention and sparsity conditions, with a position-dependent transformation of logits.
result Scale-invariant attention scheme improves validation loss and long-context retrieval.
New algorithm clusters LLM inputs efficiently with quality control.
problem Scaling LLM inference cost and latency for millions of users.
method Two-stage algorithm using Mini-batch K-Means and greedy selection of representatives.
result Enforces similarity and attribute guardrails exactly, scales to tens of millions of samples.
This paper uses LLMs to generate synthetic data to improve classification accuracy in imbalanced datasets.
problem Imbalanced classification and spurious correlation in data science.
method Develops novel theoretical foundations and uses transformer models to generate synthetic data.
result Transformer models can generate high-quality synthetic data to improve classification accuracy.
New framework selects high-quality pretraining data without training LLMs.
problem Slow progress in understanding pretraining data due to costly experiments.
method Statistical framework based on perplexity-benchmark correlations.
result Approach outperforms existing methods on multiple benchmarks.
We show LLMs can be locally linear, enabling better control of activations.
problem Suboptimal control of LLM activations during generation.
method Model LLM inference as a linear dynamical system, compute feedback controllers using Jacobians, and adapt classical control theory.
result Robust, fine-grained control of LLM activations across models and tasks.
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.
LLMs struggle to outperform markets over long periods and diverse stocks.
problem Overstated effectiveness of LLM-based investing strategies due to biases.
method FINSABER framework for systematic backtests over two decades and 100+ symbols.
result Previously reported LLM advantages deteriorate significantly under broader evaluation.
G1 uses RL to enhance LLMs' graph reasoning, improving performance on diverse tasks.
problem Limited graph reasoning abilities of LLMs, especially in synthetic graph-theoretic tasks.
method Curated synthetic graph dataset, RL training on LLMs.
result Significant improvements in graph reasoning, zero-shot generalization to unseen tasks.
New framework improves LLM performance by avoiding forgetting during sequential training stages.
problem Forgetting during sequential training stages of LLMs.
method Proposes a joint post-training framework with theoretical convergence guarantees.
result Empirically outperforms sequential post-training framework by up to 23%.
TradeExpert uses a mix of LLMs to predict stock movements.
problem Synthesizing insights from diverse financial data sources.
method A mix of four specialized LLMs analyzing different data types, with a General Expert LLM synthesizing the insights.
result TradeExpert outperforms existing benchmarks in stock movement prediction.
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.
FinGPT democratizes financial data for LLMs, enabling innovation.
problem Limited financial text datasets and disparities between general and financial text data.
method Automates collection and curation of real-time financial data from diverse Internet sources, fine-tuning with RLSP and LoRA.
result Democratizes access to financial data for LLMs, enabling innovation.
LOLA uses LLMs to optimize content delivery, outperforming traditional methods.
problem Identifying the most engaging headlines for user engagement.
method LOLA integrates LLMs with adaptive experimentation to optimize content delivery.
result LOLA outperforms traditional methods in optimizing user engagement.
LLMs struggle to optimize hyperparameters efficiently, but hybrid methods can improve performance.
problem Optimizing hyperparameters of small language models using LLMs.
method Comparison of classical HPO algorithms and LLM-based methods, introducing Centaur hybrid approach.
result Hybrid Centaur approach achieves best results, outperforming classical and pure LLM methods.
Duel-Evolve uses LLM self-preferences for test-time optimization of discrete outputs.
problem Optimizing LLM outputs at test time with limited or unreliable scalar rewards.
method Duel-Evolve uses pairwise comparisons from the LLM to guide optimization, aggregating them via a Bayesian Bradley-Terry model.
result Achieves significant improvement over existing methods in accuracy.
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.
A new framework evaluates large language models efficiently and accurately.
problem Evaluation of large language models is challenging due to stochasticity and heterogeneity of benchmarks.
method Interpretable and scalable framework based on Item Response Theory (IRT) and majorization-minimization principle.
result Our method achieves superior scalability and interpretability compared to existing approaches.
Improved unlearning in LLMs using contrastive decoding.
problem Removing specific information from LLMs while preserving overall performance.
method Inference-time unlearning algorithm using contrastive decoding with two auxiliary models.
result Notable gains in forget quality and retained performance compared to prior approaches.
This study examines how the size and alignment of pretraining data affect the performance of large language models on downstream tasks.
problem Understanding how the size and alignment of pretraining data impact the performance of large language models on downstream tasks.
method Investigated the scaling behavior of large language models in a transfer learning setting, focusing on machine translation tasks.
result The size of the finetuning dataset and the distribution alignment between pretraining and downstream data significantly influence the scaling behavior of downstream performance.
This paper proposes a method to use LLMs as auxiliary evaluators in place of human judges.
problem The need for cost-effective and scalable evaluation of AI systems.
method Formulates a two-stage sampling design with LLM evaluations and human ratings, using a doubly robust estimator.
result Proposes a method to determine optimal sample sizes for human and LLM ratings.
New framework reveals thermodynamic principles for LLM training.
problem Understanding the training dynamics of large language models.
method Introducing Neural Thermodynamic Laws (NTL) under river-valley loss landscape assumptions.
result Key thermodynamic quantities and principles naturally emerge in LLM training.
Compressed LLM embeddings improve noisy regression tasks without overfitting.
problem Noisy regression tasks with high signal-to-noise ratios.
method Comparison of embedding compression techniques using autoencoder hidden representations.
result Compression improves performance on noisy tasks like financial return prediction.
This paper evaluates LLMs on large graph property estimation tasks.
problem Limited context length of LLMs limits their evaluation on large graphs.
method Developed EstGraph dataset and introduced four tasks for LLMs to estimate large graph properties.
result LLMs perform better on graph property estimation tasks when provided with context-rich prompts based on random walks.
This work extends the scaling law to multiple and kernel regression, challenging traditional machine learning principles.
problem Challenging traditional machine learning wisdom with scaling law in large practical models.
method Demonstrates the scaling law in multiple and kernel regression settings.
result The scaling law extends to multiple and kernel regression, providing deeper insights into LLMs.
Large language models correlate in errors, even with different architectures and providers.
problem Lack of empirical evidence on whether different large language models differ meaningfully.
method Empirical evaluation of over 350 large language models using two leaderboards and a resume-screening task.
result Large language models have substantial correlation in errors, even with distinct architectures and providers.
A new algorithm optimizes softmax units in large language models.
problem Efficiently computing gradients for large-scale language models.
method Zero-th Order method for approximating gradients.
result The algorithm converges and efficiently computes gradients.
Fractured Sampling improves LLM reasoning efficiency by truncating CoT trajectories.
problem Efficiently scaling reasoning in large language models with limited tokens.
method Integrating truncated Chain-of-Thought (CoT) with Fractured Sampling across multiple dimensions.
result Fractured Sampling achieves superior accuracy-cost trade-offs compared to full CoT.
New analysis shows LLMs don't follow Bayesian inference in ICL.
problem Does in-context learning in LLMs follow Bayesian inference?
method Analyzes ICL through the martingale property, a requirement for Bayesian inference.
result Violations of the martingale property show LLMs don't follow Bayesian inference.
Novel method diagnoses large language models' reasoning abilities.
problem Fine-grained evaluation of large language models' reasoning abilities.
method Adapting cognitive diagnosis models to LLMs, estimating mastery profiles and Q-matrix, incorporating textual information.
result Accurate parameter recovery and insights into LLMs' capabilities.
Influence functions help study large language model generalization, revealing surprising decay patterns.
problem Understanding and mitigating risks in large language models (LLMs).
method Eigenvalue-corrected Kronecker-Factored Approximation (EK-FAC) to scale influence functions to LLMs.
result Influences decay to near-zero when key phrases order is flipped, revealing a surprising limitation.