Soft Contamination Means Benchmarks Test Shallow Generalization
This addresses the issue of overestimated generalization in LLM benchmarks for researchers and practitioners, revealing that benchmark improvements may be incremental due to data contamination rather than genuine capability gains.
The study investigated the problem of soft contamination in LLM training data, where semantic duplicates of benchmark test data bias performance estimates, finding that 78% of CodeForces and 50% of ZebraLogic problems had such duplicates, which improved benchmark performance and confounded recent gains.
If LLM training data is polluted with benchmark test data, then benchmark performance gives biased estimates of out-of-distribution (OOD) generalization. Typical decontamination filters use n-gram matching which fail to detect semantic duplicates: sentences with equivalent (or near-equivalent) content that are not close in string space. We study this soft contamination of training data by semantic duplicates. Among other experiments, we embed the Olmo3 training corpus and find that: 1) contamination remains widespread, e.g. we find semantic duplicates for 78% of CodeForces and exact duplicates for 50% of ZebraLogic problems; 2) including semantic duplicates of benchmark data in training does improve benchmark performance; and 3) when finetuning on duplicates of benchmark datapoints, performance also improves on truly-held-out datapoints from the same benchmark. We argue that recent benchmark gains are thus confounded: the prevalence of soft contamination means gains reflect both genuine capability improvements and the accumulation of test data and effective test data in growing training corpora.