We argue that the machine learning value chain is structurally unsustainable due to an economic data processing inequality: each state in the data cycle from inputs to model weights to synthetic outputs refines technical signal but strips economic equity from data generators. We show, by analyzing seventy-three public data deals, that the majority of value accrues to aggregators, with documented creator royalties rounding to zero and widespread opacity of deal terms. This is not just an economic welfare concern: as data and its derivatives become economic assets, the feedback loop that sustains current learning algorithms is at risk. We identify three structural faults - missing provenance, asymmetric bargaining power, and non-dynamic pricing - as the operational machinery of this inequality. In our analysis, we trace these problems along the machine learning value chain and propose an Equitable Data-Value Exchange (EDVEX) Framework to enable a minimal market that benefits all participants. Finally, we outline research directions where our community can make concrete contributions to data deals and contextualize our position with related and orthogonal viewpoints.

In post-training for reasoning Large Language Models (LLMs), the current state of practice trains LLMs in two independent stages: Supervised Fine-Tuning (SFT) and Reinforcement Learning with Verifiable Rewards (RLVR, shortened as "RL" below). In this work, we challenge whether high SFT scores translate to improved performance after RL. We provide extensive counter-examples where this is not true. We find high SFT scores can be biased toward simpler or more homogeneous data and are not reliably predictive of subsequent RL gains or scaled-up post-training effectiveness. In some cases, RL training on models with improved SFT performance could lead to substantially worse outcome compared to RL on the base model without SFT. We study alternative metrics and identify generalization loss on held-out reasoning examples and Pass@large k performance to provide strong proxies for the RL outcome. We trained hundreds of models up to 12B-parameter with SFT and RLVR via GRPO and ran extensive evaluations on 7 math benchmarks with up to 256 repetitions, spending $>$1M GPU hours. Experiments include models from Llama3, Mistral-Nemo, Qwen3 and multiple state-of-the-art SFT/RL datasets. Compared to directly predicting from pre-RL performance, prediction based on generalization loss and Pass@large k achieves substantial higher precision, improving $R^2$ coefficient and Spearman's rank correlation coefficient by up to 0.5 (2x). This provides strong utility for broad use cases. For example, in most experiments, we find SFT training on unique examples for a one epoch underperforms training on half examples for two epochs, either after SFT or SFT-then-RL; With the same SFT budget, training only on short examples may lead to better SFT performance, though, it often leads to worse outcome after RL compared to training on examples with varying lengths. This work develops an enhanced evaluation tool that will be open-sourced.

Curating training data is among the most consequential yet labor-intensive parts of modern AI development: practitioners iteratively propose, implement, evaluate, and revise data policies against noisy benchmark feedback. We ask whether generalist coding agents can automate this data-curation loop. We introduce *Curation-Bench*, an agent-centric benchmark that fixes the model, training recipe, and evaluation suite while giving agents command-line access to inspect data, implement policies, submit them to a fixed training/evaluation pipeline, and revise. In a vision-language instruction-tuning instantiation, out-of-the-box agents reach strong published data-selection baselines within ten iterations. However, trajectory analysis reveals a persistent *execution-research gap*: agents mainly tune local policy variants rather than explore new policy families, even when given strategy guides and paper references. Scaffolds requiring each iteration to cite, instantiate, and adapt a prior method shift agents toward method-guided exploration. The scaffolded agent autonomously composes – without human design input – a data-selection policy that outperforms strong published baselines at one-tenth their data budget. Overall, current agents can run the curation loop, but reliable data research requires scaffolded method adaptation, not open-ended prompting alone. Code and benchmark are open-sourced.

Skills are a natural unit for describing what a language model can do and how its behavior can be changed. However, existing characterizations rely on human-written taxonomies, textual descriptions, or manual profiling pipelines–all external hypotheses about what matters that need not align with the model's internal representations. We argue that when the goal is to intervene on model behavior, skill characterization should be *model-native*: grounded in the model's own representations rather than imposed through external ontologies. We instantiate this view by recovering a compact orthogonal basis from sequence-level activations. The resulting basis is semantically interpretable but need not correspond to any predefined human ontology; instead, it captures axes of behavioral variation that the model itself organizes around. We validate this characterization on reasoning post-training, using the recovered basis for both SFT data selection and inference-time steering. We develop lightweight proxy interventions to identify which directions are most useful for a given model. Across Llama3-8B and Qwen2.5-3B, selecting data along those directions improves Pass@1 by up to 20

Domain reweighting can improve sample efficiency and downstream generalization, but data-mixture optimization for multimodal midtraining remains largely unexplored. Current multimodal training recipes tune mixtures along a single dimension, typically data format or task type. We introduce MixAtlas, a method that produces benchmark-targeted data recipes that can be inspected, adapted, and transferred to new corpora. MixAtlas decomposes the training corpus along two axes: image concepts (10 visual-domain clusters discovered via CLIP embeddings) and task supervision (5 objective types including captioning, OCR, grounding, detection, and VQA). Using small proxy models (Qwen2-0.5B) paired with a Gaussian-process surrogate and GP-UCB acquisition, MixAtlas searches the resulting mixture space with the same proxy budget as regression-based baselines but finds better-performing mixtures. We evaluate on 10 benchmarks spanning visual understanding, document reasoning, and multimodal reasoning. On Qwen2-7B, optimized mixtures improve average performance by 8.5

Neural scaling laws are widely used for performance projection and resource planning, yet their sensitivity to data quality interventions remains poorly understood. We present an empirical study of how interventions—deduplication, heuristic filtering, and LLM-guided rewriting—reshape scaling behavior in large language model training. Using QualityPajama, a suite of 23 systematically filtered and synthetic datasets, we train over 2,000 models (100M–8B parameters, 100M–200B tokens) to measure how data quality affects scaling-law parameters and compute-optimal design decisions. Our results show that data interventions reshape scaling dynamics in non-trivial ways not captured by current theory, simultaneously moving exponents, coefficients, and constants in conflicting directions that exert opposing forces on loss. For example, an intervention may improve constants but hurt the exponents. Strategies that appear optimal at small scale can reverse at larger scale, and compute-optimal token–parameter ratios can vary by orders of magnitude depending on the intervention. These findings demonstrate that data curation and scaling strategy are deeply intertwined, and that evaluating interventions only at fixed scales can lead to misleading conclusions. We recommend evaluating interventions through their full scaling trajectories using scaling law projections.

The performance of an LLM agent depends on the scaffold around a frozen model. A common way to improve that scaffold is to use a coding agent as an optimizer: it reads current scores and traces and iteratively edits the source, producing a new candidate each round. Each edit is chosen according to a belief about how the environment will respond: what went wrong, and which change should help. That belief is typically implicit. It lives in the coding agent's reasoning on the current call, or remains latent in its parameters, rather than as something written down. Later calls therefore see scores and traces, but they do not use that belief. We introduce Belief-Calibrated Optimization (BCO), a method that writes that belief down as a persistent in-context document and continually revises that document as new candidates are evaluated. The resulting document is a world model: the current account of how the environment responds to edits. Added to an otherwise standard loop, BCO reaches a higher train passrate than a matched control that lacks only the world model, on five benchmarks spanning memory QA, tool-use QA, code-as-action app agents, and terminal agents. The gap remains on every held-out split, which is not used to select the candidate. After a target-model swap, in which the frozen model is replaced and the scaffold is not, the selected BCO scaffold leads on the tasks we test, except where context-window overruns leave it unfinished. An offline ablation then asks whether that gap comes from what the world model says. A fresh predictor given the accumulated document forecasts how the environment will respond more accurately than predictors given either no document or a same-form copy whose content has been falsified. The comparison indicates that the document carries reusable information in its content, not only in its form.

As Machine Learning (ML) systems continue to grow, the demand for relevant and comprehensive datasets becomes imperative. There is limited study on the challenges of data acquisition due to ad-hoc processes and lack of consistent methodologies. We first present an investigation of current data marketplaces, revealing lack of platforms offering detailed information about datasets, transparent pricing, standardized data formats. With the objective of inciting participation from the data-centric AI community, we then introduce the DAM challenge, a benchmark to model the interaction between the data providers and acquirers. The benchmark was released as a part of DataPerf. Our evaluation of the submitted strategies underlines the need for effective data acquisition strategies in ML.

Training data plays a crucial role in Large Language Models (LLM) scaling, yet high quality data is of limited supply. Synthetic data techniques offer a potential path toward sidestepping these limitations.We conduct a large-scale empirical investigation (>1000 LLMs with >100k GPU hours) using a unified protocol and scaling laws, comparing natural web data, diverse synthetic types (rephrased text, generated textbooks), and mixtures of natural and synthetic data. Specifically, we found pre-training on rephrased synthetic data alone is not faster than pre-training on natural web texts; while pre-training on 1/3 rephrased synthetic data mixed with 2/3 natural web texts can speed up 5-10x (to reach the same validation loss) at larger data budgets. Pre-training on textbook-style synthetic data alone results in notably higher loss on many downstream domains especially at small data budgets. “Good” ratios of synthetic data in training data mixtures depend on the model size and data budget, empirically converging to ~30% for rephrased synthetic data. Larger generator models do not necessarily yield better pre-training data than ~8B-param models. These results contribute mixed evidence on “model collapse” during large-scale single-round (n=1) model training on synthetic data–training on rephrased synthetic data shows no degradation in performance in foreseeable scales whereas training on mixtures of textbook-style pure-generated synthetic data shows patterns predicted by “model collapse”. Our work demystifies synthetic data in pre-training, validates its conditional benefits, and offers practical guidance.

The computational burden and inherent redundancy of large-scale datasets challenge the training of contemporary machine learning models. Data pruning offers a solution by selecting smaller, informative subsets, yet existing methods struggle: density-based approaches can be task-agnostic, while model-based techniques may introduce redundancy or prove computationally prohibitive. We introduce Adaptive De-Duplication (AdaDeDup), a novel hybrid framework that synergistically integrates density-based pruning with model-informed feedback in a cluster-adaptive manner. AdaDeDup first partitions data and applies an initial density-based pruning. It then employs a proxy model to evaluate the impact of this initial pruning within each cluster by comparing losses on kept versus pruned samples. This task-aware signal adaptively adjusts cluster-specific pruning thresholds, enabling more aggressive pruning in redundant clusters while preserving critical data in informative ones. Extensive experiments on large-scale object detection benchmarks (Waymo, COCO, nuScenes) using standard models (BEVFormer, Faster R-CNN) demonstrate AdaDeDup's advantages. It significantly outperforms prominent baselines, substantially reduces performance degradation (e.g., over 54

Fact tracing seeks to identify specific training examples that serve as the knowledge source for a given query. Existing approaches to fact tracing rely on assessing the similarity between each training sample and the query along a certain dimension, such as lexical similarity, gradient, or embedding space. However, these methods fall short of effectively distinguishing between samples that are merely relevant and those that actually provide supportive evidence for the information sought by the query. This limitation often results in suboptimal effectiveness. Moreover, these approaches necessitate the examination of the similarity of individual training points for each query, imposing significant computational demands and creating a substantial barrier for practical applications. This paper introduces FASTTRACK, a novel approach that harnesses the capabilities of Large Language Models (LLMs) to validate supportive evidence for queries and at the same time clusters the training database towards a reduced extent for LLMs to trace facts. Our experiments show that FASTTRACK substantially outperforms existing methods in both accuracy and efficiency, achieving more than 100% improvement in F1 score over the state-of-the-art methods while being x33 faster than TracIn.

Large-scale black-box models have become ubiquitous across numerous applications. Understanding the influence of individual training data sources on predictions made by these models is crucial for improving their trustworthiness. Current influence estimation techniques involve computing gradients for every training point or repeated training on different subsets. These approaches face obvious computational challenges when scaled up to large datasets and models. In this paper, we introduce and explore the Mirrored Influence Hypothesis, highlighting a reciprocal nature of influence between training and test data. Specifically, it suggests that evaluating the influence of training data on test predictions can be reformulated as an equivalent, yet inverse problem: assessing how the predictions for training samples would be altered if the model were trained on specific test samples. Through both empirical and theoretical validations, we demonstrate the wide applicability of our hypothesis. Inspired by this, we introduce a new method for estimating the influence of training data, which requires calculating gradients for specific test samples, paired with a forward pass for each training point. This approach can capitalize on the common asymmetry in scenarios where the number of test samples under concurrent examination is much smaller than the scale of the training dataset, thus gaining a significant improvement in efficiency compared to existing approaches. We demonstrate the applicability of our method across a range of scenarios, including data attribution in diffusion models, data leakage detection, analysis of memorization, mislabeled data detection, and tracing behavior in language models.

Machine Learning models have shown susceptibility to various privacy attacks, with model inversion (MI) attacks posing a significant threat. Current defense techniques are mostly \emph{model-centric}, involving modifying model training or inference. However, these approaches require model trainers' cooperation, are computationally expensive, and often result in a significant privacy-utility tradeoff. To address these limitations, we propose a novel \emph{data-centric} approach to mitigate MI attacks. Compared to traditional model-centric techniques, our approach offers the unique advantage of enabling each individual user to control their data's privacy risk, aligning with findings from a Cisco survey that only a minority actively seek privacy protection. Specifically, we introduce several privacy-focused data augmentations that modify the private data uploaded to the model trainer. These augmentations shape the resulting model's loss landscape, making it challenging for attackers to generate private target samples. Additionally, we provide theoretical analysis to explain why such augmentations can reduce the risk of model inversion. We evaluate our approach against state-of-the-art MI attacks and demonstrate its effectiveness and robustness across various model architectures and datasets. Specifically, in standard face recognition benchmarks, we reduce face reconstruction success rates to $\leq5\%$, while maintaining high utility with only a 2\% classification accuracy drop, significantly surpassing state-of-the-art model-centric defenses. This is the first study to propose a data-centric approach for mitigating model inversion attacks, showing promising potential for decentralized privacy protection.

Domain reweighting is an emerging research area aimed at adjusting the relative weights of different data sources to improve the effectiveness and efficiency of LLM pre-training. We show that data mixtures that perform well at smaller scales may not retain their advantage at larger scales, challenging the existing practice of determining competitive mixtures in small-scale experiments and directly applying them at much larger scales. To address this, we propose AutoScale, a two-stage, scale-aware data composition framework. First, AutoScale fits a parametric model that predicts the model's loss under different data compositions, then uses it to find an approximate best allocation at smaller, more manageable budgets. Next, leveraging a novel theoretical analysis of how optimal compositions evolve with scale, AutoScale extrapolates that composition to larger budgets without further retraining. Empirically, AutoScale accelerates convergence and improves downstream performance. For instance, when pre-training GPT-2 Large, it achieves a 28 faster perplexity reduction than baselines and up to a 38 unweighted training, while yielding best-average results on various downstream tasks. Overall, our findings illustrate how domain importance shifts with training scale, underscoring the need for scale-dependent data curation in LLM training. Our code is open-sourced.

Fact tracing seeks to identify specific training examples that serve as the knowledge source for a given query. Existing approaches to fact tracing rely on assessing the similarity between each training sample and the query along a certain dimension, such as lexical similarity, gradient, or embedding space. However, these methods fall short of effectively distinguishing between samples that are merely relevant and those that actually provide supportive evidence for the information sought by the query. This limitation often results in suboptimal effectiveness. Moreover, these approaches necessitate the examination of the similarity of individual training points for each query, imposing significant computational demands and creating a substantial barrier for practical applications. This paper introduces FASTTRACK, a novel approach that harnesses the capabilities of Large Language Models (LLMs) to validate supportive evidence for queries and at the same time clusters the training database towards a reduced extent for LLMs to trace facts. Our experiments show that FASTTRACK substantially outperforms existing methods in both accuracy and efficiency, achieving more than 100\% improvement in F1 score over the state-of-the-art methods while being X33 faster than \texttt{TracIn}.

Machine learning (ML) models are widely used but vulnerable to privacy attacks, such as model inversion. Current defense techniques are mostly model-centric, involving modifying the model training or inference process. However, these approaches require model trainers' cooperation, are computationally expensive, and often result in a significant privacy-utility tradeoff. To address these limitations, this paper proposes a novel data-centric approach to mitigate model inversion attacks. Compared to traditional model-centric techniques, our approach offers the unique advantage of decentralization, enabling individual users to control their data's privacy risk. Specifically, we introduce several privacy-focused data augmentations that modify the private data uploaded to the model trainer. These augmentations will shape the resulting model's loss landscape, making it challenging for attackers to generate private target samples. Additionally, we provide theoretical analysis to support our approach and explain how data augmentation can reduce the risk of model inversion. We evaluate our approach against state-of-the-art model inversion attacks and demonstrate its effectiveness and robustness across various model architectures and datasets. Specifically, in standard face recognition benchmarks, we reduce face reconstruction success rates to less or equal to 5%, while maintaining high utility with only a 2% classification accuracy drop, surpassing state-of-the-art model-centric defenses by 90%. This is the first study to propose a data-centric approach for mitigating model inversion attacks, showing potential for decentralized privacy protection.

This work focuses on leveraging and selecting from vast, unlabeled, open data to \emph{pre-fine-tune} a pre-trained language model. The goal is to minimize the need for costly domain-specific data for subsequent fine-tuning while achieving desired performance levels. While many data selection algorithms have been designed for small-scale applications, rendering them unsuitable for our context, some emerging methods do cater to language data scales. However, they often prioritize data that aligns with the target distribution. While this strategy may be effective when training a model from scratch, it can yield limited results when the model has already been pre-trained on a different distribution. Differing from prior work, our key idea is to select data that nudges the pre-training distribution closer to the target distribution. We show the optimality of this approach for fine-tuning tasks under certain conditions. We demonstrate the efficacy of our methodology across a diverse array of tasks, showing that it consistently surpasses other selection methods. Moreover, our proposed method is significantly faster than existing techniques, scaling to millions of samples within a single GPU hour. Our code is open-sourced \footnote{Code repository: \url{https://anonymous.4open.science/r/DV4LLM-D761/}}. While fine-tuning offers significant potential for enhancing performance across diverse tasks, its associated costs often limit its widespread adoption; with this work, we hope to lay the groundwork for cost-effective fine-tuning, making its benefits more accessible.

Traditionally, data selection has been studied in settings where all samples from prospective sources are fully revealed to a machine learning developer. However, in practical data exchange scenarios, data providers often reveal only a limited subset of samples before an acquisition decision is made. Recently, there have been efforts to fit scaling laws that predict model performance at any size and data source composition using the limited available samples. However, these scaling functions are black-box, computationally expensive to fit, highly susceptible to overfitting, or/and difficult to optimize for data selection. This paper proposes a framework called , which predicts model performance and supports data selection decisions based on partial samples of prospective data sources. Our approach distinguishes itself from existing work by introducing a novel *two-stage* performance inference process. In the first stage, we leverage the Optimal Transport distance to predict the model's performance for any data mixture ratio within the range of disclosed data sizes. In the second stage, we extrapolate the performance to larger undisclosed data sizes based on a novel parameter-free mapping technique inspired by neural scaling laws. We further derive an efficient gradient-based method to select data sources based on the projected model performance. Evaluation over a diverse range of applications demonstrates that significantly improves existing performance scaling approaches in terms of both the accuracy of performance inference and the computation costs associated with constructing the performance predictor. Also, outperforms by a wide margin in data selection effectiveness compared to a range of other off-the-shelf solutions.

Traditionally, data valuation (DV) is posed as a problem of equitably splitting the validation performance of a learning algorithm among the training data. As a result, the calculated data values depend on many design choices of the underlying learning algorithm. However, this dependence is undesirable for many DV use cases, such as setting priorities over different data sources in a data acquisition process and informing pricing mechanisms in a data marketplace. In these scenarios, data needs to be valued before the actual analysis and the choice of the learning algorithm is still undetermined then. Another side-effect of the dependence is that to assess the value of individual points, one needs to re-run the learning algorithm with and without a point, which incurs a large computation burden. This work leapfrogs over the current limits of data valuation methods by introducing a new framework that can value training data in a way that is oblivious to the downstream learning algorithm. Our main results are as follows. (1) We develop a proxy for the validation performance associated with a training set based on a non-conventional class-wise Wasserstein distance between training and validation sets. We show that the distance characterizes the upper bound of the validation performance for any given model under certain Lipschitz conditions. (2) We develop a novel method to value individual data based on the sensitivity analysis of the class-wise Wasserstein distance. Importantly, these values can be directly obtained for free from the output of off-the-shelf optimization solvers when computing the distance. (3) We evaluate our new data valuation framework over various use cases related to detecting low-quality data and show that, surprisingly, the learning-agnostic feature of our framework enables a significant improvement over SOTA performance while being orders of magnitude faster.