Researcher — paper.dinq.me

作者位次 #3/5ICLR 2026Poster· Applications / Language, Speech and Dialog

Towards Reliable Benchmarking: A Contamination Free, Controllable Evaluation Framework for Multi-step LLM Function Calling

Seiji Maekawa · Jackson Hassell · Pouya Pezeshkpour · Tom Mitchell · Estevam Hruschka

评测与基准多模态OpenReview

作者位次 #2/3ACL 2026Poster· Natural Language Processing

From Proof to Program: Characterizing Tool-Induced Reasoning Hallucinations in Large Language Models

Farima Fatahi Bayat · Pouya Pezeshkpour · Estevam Hruschka

摘要

Tool-augmented Language Models (TaLMs) can invoke external tools to solve problems beyond their parametric capacity. However, it remains unclear whether these tool-enabled gains reflect trustworthy reasoning. Focusing on the Code Interpreter tool, we show that even when tools are selected and exe...

推理对齐与偏好评测与基准OpenReview 项目页

Pouya Pezeshkpour

获取完整研究档案

联系方式

2026 顶会论文

Towards Reliable Benchmarking: A Contamination Free, Controllable Evaluation Framework for Multi-step LLM Function Calling

From Proof to Program: Characterizing Tool-Induced Reasoning Hallucinations in Large Language Models

常见合作者