Portrait of Paolo Cudrano is unavailable

Paolo Cudrano

Alumni
Website
Google Scholar

Publications

Influence Functions for Efficient Data Selection in Reasoning
Prateek Humane
Paolo Cudrano
Daniel Z. Kaplan
Matteo Matteucci
Supriyo Chakraborty
Irina Rish
Fine-tuning large language models (LLMs) on chain-of-thought (CoT) data shows that a small amount of high-quality data can outperform massiv… (see more)e datasets. Yet, what constitutes "quality" remains ill-defined. Existing reasoning methods rely on indirect heuristics such as problem difficulty or trace length, while instruction-tuning has explored a broader range of automated selection strategies, but rarely in the context of reasoning. We propose to define reasoning data quality using influence functions, which measure the causal effect of individual CoT examples on downstream accuracy, and introduce influence-based pruning, which consistently outperforms perplexity and embedding-based baselines on math reasoning within a model family.
2025-10-15
ER @ Neural Information Processing Systems (accepted)
doi.org
openreview.net