An empirical study distills a taxonomy of human factual errors from newspaper corrections and shows LLMs achieve only 52% F1 on detection.
Understanding Iterative Revision from Human-Written Text
3 Pith papers cite this work. Polarity classification is still indexing.
3
Pith papers citing it
fields
cs.CL 3years
2026 3verdicts
UNVERDICTED 3representative citing papers
Reinforcement learning with a multi-part reward teaches LLMs to output independent, meaning-preserving sentence edits that raise argument appropriateness close to full rewriting.
Pilot experiment shows limited LLM access maintains higher student ownership and strategic use than unlimited access, with no difference in essay quality.
citing papers explorer
-
An Empirical Analysis of Factual Errors in Human-Written Text and its Application
An empirical study distills a taxonomy of human factual errors from newspaper corrections and shows LLMs achieve only 52% F1 on detection.
-
Teaching LLMs Human-Like Editing of Inappropriate Argumentation via Reinforcement Learning
Reinforcement learning with a multi-part reward teaches LLMs to output independent, meaning-preserving sentence edits that raise argument appropriateness close to full rewriting.
-
Effects of Varying LLM Access on Essay Writing Behavior
Pilot experiment shows limited LLM access maintains higher student ownership and strategic use than unlimited access, with no difference in essay quality.