SCOP tests five comprehension skills separately and shows LLMs are far better at local than global comprehension, often answering correctly without identifying the right supporting evidence.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
SCOP: Evaluating the Comprehension Process of Large Language Models from a Cognitive View
SCOP tests five comprehension skills separately and shows LLMs are far better at local than global comprehension, often answering correctly without identifying the right supporting evidence.