LLMs cannot spot math errors, even when allowed to peek into the solution

Documenti analoghi

• Can LLMs Reliably Simulate Real Students' Abilities in Mathematics and Reading Comprehension?
  di: Srivatsa, KV Aditya, et al.
  Pubblicazione: (2025)
• Simulating LLM-to-LLM Tutoring for Multilingual Math Feedback
  di: Tonga, Junior Cedric, et al.
  Pubblicazione: (2025)
• SelectLLM: Query-Aware Efficient Selection Algorithm for Large Language Models
  di: Maurya, Kaushal Kumar, et al.
  Pubblicazione: (2024)
• Harnessing the Power of Multiple Minds: Lessons Learned from LLM Routing
  di: Srivatsa, KV Aditya, et al.
  Pubblicazione: (2024)
• Findings of the BEA 2025 Shared Task on Pedagogical Ability Assessment of AI-powered Tutors
  di: Kochmar, Ekaterina, et al.
  Pubblicazione: (2025)