Papers by Stefano Civelli

1 papers
A Shared Geometry of Difficulty in Multilingual Language Models (2026.acl-short)

Copied to clipboard

Challenge: Large language models encode problem difficulty as an internal signal that can be linearly decoded from their residuals.
Approach: They train linear probes on the AMC subset of the Easy2Hard benchmark, translated into 21 languages, and found difficulty-related signals emerge at two distinct stages of the model internals.
Outcome: The results show that difficulty-related signals emerge at two distinct stages of the model internals, corresponding to shallow (early-layers) and deep (later-layer) representations, that exhibit functionally different behaviors.

What is GenGO?

GenGO is an NLP powered publication search system. It currenctly indexes 30k+ papers from ACL Anthology, and implements multi-aspect summarization, semantic search, and more!

Information

About
Limitations