# Benchmarks Are Not That Out of Distribution: Word Overlap Predicts Performance

arXiv: https://arxiv.org/abs/2602.10657 · HTML: https://mlsanity.com/paper/2602.10657/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Builds on 31 of its 48 references that are in the map.
- First posted on arXiv 11 Feb 2026.

## Builds on (most cited first, 30 of 31)

- [Llama 2: Open Foundation and Fine-Tuned Chat Models](https://mlsanity.com/paper/2307.09288/) (2023)
- [Measuring Massive Multitask Language Understanding](https://mlsanity.com/paper/2009.03300/) (2020)
- [Think you have Solved Question Answering? Try ARC, the AI2 Reasoning Challenge](https://mlsanity.com/paper/1803.05457/) (2018)
- [HellaSwag: Can a Machine Really Finish Your Sentence?](https://mlsanity.com/paper/1905.07830/) (2019)
- [Neural Machine Translation of Rare Words with Subword Units](https://mlsanity.com/paper/1508.07909/) (2015)
- [The FineWeb Datasets: Decanting the Web for the Finest Text Data at Scale](https://mlsanity.com/paper/2406.17557/) (2024)
- [The LAMBADA dataset: Word prediction requiring a broad discourse context](https://mlsanity.com/paper/1606.06031/) (2016)
- [Crowdsourcing Multiple Choice Science Questions](https://mlsanity.com/paper/1707.06209/) (2017)
- [MathQA: Towards Interpretable Math Word Problem Solving with Operation-Based Formalisms](https://mlsanity.com/paper/1905.13319/) (2019)
- [BLiMP: The Benchmark of Linguistic Minimal Pairs for English](https://mlsanity.com/paper/1912.00582/) (2019)
- [Documenting Large Webtext Corpora: A Case Study on the Colossal Clean Crawled Corpus](https://mlsanity.com/paper/2104.08758/) (2021)
- [Dolma: an Open Corpus of Three Trillion Tokens for Language Model Pretraining Research](https://mlsanity.com/paper/2402.00159/) (2024)
- [DataComp-LM: In search of the next generation of training sets for language models](https://mlsanity.com/paper/2406.11794/) (2024)
- [Language Modeling Is Compression](https://mlsanity.com/paper/2309.10668/) (2023)
- [FineWeb2: One Pipeline to Scale Them All -- Adapting Pre-Training Data Processing to Every Language](https://mlsanity.com/paper/2506.20920/) (2025)
- [Scaling Laws with Vocabulary: Larger Models Deserve Larger Vocabularies](https://mlsanity.com/paper/2407.13623/) (2024)
- [Language models scale reliably with over-training and on downstream tasks](https://mlsanity.com/paper/2403.08540/) (2024)
- [Stable LM 2 1.6B Technical Report](https://mlsanity.com/paper/2402.17834/) (2024)
- [Tokenization Is More Than Compression](https://mlsanity.com/paper/2402.18376/) (2024)
- [Heavy-Tailed Class Imbalance and Why Adam Outperforms Gradient Descent on Language Models](https://mlsanity.com/paper/2402.19449/) (2024)
- [SlimPajama-DC: Understanding Data Combinations for LLM Training](https://mlsanity.com/paper/2309.10818/) (2023)
- [The No Free Lunch Theorem, Kolmogorov Complexity, and the Role of Inductive Biases in Machine Learning](https://mlsanity.com/paper/2304.05366/) (2023)
- [Compression Represents Intelligence Linearly](https://mlsanity.com/paper/2404.09937/) (2024)
- [DataDecide: How to Predict Best Pretraining Data with Small Experiments](https://mlsanity.com/paper/2504.11393/) (2025)
- [Ultra-FineWeb: Efficient Data Filtering and Verification for High-Quality LLM Training Data](https://mlsanity.com/paper/2505.05427/) (2025)
- [A Definition of AGI](https://mlsanity.com/paper/2510.18212/) (2025)
- [Understanding and Mitigating Tokenization Bias in Language Models](https://mlsanity.com/paper/2406.16829/) (2024)
- [Causal Estimation of Tokenisation Bias](https://mlsanity.com/paper/2506.03149/) (2025)
- [Compression via Pre-trained Transformers: A Study on Byte-Level Multimodal Data](https://mlsanity.com/paper/2410.05078/) (2024)
- [Exploiting Vocabulary Frequency Imbalance in Language Model Pre-training](https://mlsanity.com/paper/2508.15390/) (2025)

## Built on by (most cited first, 1 of 1)

- [Benchmarking the Benchmarks: Testing the Predictive Validity of Commonsense Benchmarks](https://mlsanity.com/paper/2608.03340/) (2026)

## Abstract (arXiv)

Understanding what constitutes high-quality pre-training data remains a central question in language model training. In this work, we investigate whether benchmark performance is primarily driven by the degree of statistical pattern overlap between pre-training corpora and evaluation datasets. We measure this overlap using word-level unigram cross-entropy and word frequency statistics, and perform controlled experiments across 10 zero-shot benchmarks, 4 pre-training datasets spanning 8.5B to 60B tokens, and model sizes ranging from 400M to 3B parameters. Our results demonstrate a robust inverse relationship between word-level unigram cross-entropy and benchmark performance, suggesting that widely used benchmarks are strongly influenced by word overlap between training and evaluation data. Thus, larger pre-training subsets with similar word-level unigram cross-entropy yield improved downstream results, indicating that word frequency statistics play an additional role in shaping benchmark scores. Taken together, these results suggest that many standard benchmarks are only weakly out-of-distribution relative to pre-training corpora, so that simple word-overlap statistics predict benchmark performance.