Think Twice: Measuring the Efficiency of Eliminating Prediction Shortcuts of Question Answering Models

Documenti analoghi

• Concept-aware Data Construction Improves In-context Learning of Language Models
  di: Štefánik, Michal, et al.
  Pubblicazione: (2024)
• Can Out-of-Distribution Evaluations Uncover Reliance on Shortcuts? A Case Study in Question Answering
  di: Štefánik, Michal, et al.
  Pubblicazione: (2025)
• Self-training Language Models for Arithmetic Reasoning
  di: Kadlčík, Marek, et al.
  Pubblicazione: (2024)
• Negation: A Pink Elephant in the Large Language Models' Room?
  di: Vrabcová, Tereza, et al.
  Pubblicazione: (2025)
• Don't Think Twice! Over-Reasoning Impairs Confidence Calibration
  di: Lacombe, Romain, et al.
  Pubblicazione: (2025)