Mid-sized LLMs outperform larger models on fairness in multi-document news summarization, with entity sentiment bias proving hardest to mitigate across prompt and judge-based interventions.
argued,” “stated,
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
When Bigger Isn't Better: A Comprehensive Fairness Evaluation of Political Bias in Multi-News Summarisation
Mid-sized LLMs outperform larger models on fairness in multi-document news summarization, with entity sentiment bias proving hardest to mitigate across prompt and judge-based interventions.