paper-with-me

홈 › Papers

MMLU-Pro+: Evaluating Higher-Order Reasoning and Shortcut Learning in LLMs

2024-09-03 · Saeid Asgari Taghanaki, Aliasgahr Khani, Amir Khasahmadi

Existing benchmarks for large language models (LLMs) increasingly struggle to differentiate between top-performing models, underscoring the need for more challenging evaluation frameworks. We introduce MMLU-Pro+, an enhanced benchmark building upon MMLU-Pro to assess shortcut learning and higher-order reasoning in LLMs. By incorporating questions with multiple correct answers across diverse domains, MMLU-Pro+ tests LLMs' ability to engage in complex reasoning and resist simplistic problem-solving strategies. Our results show that MMLU-Pro+ maintains MMLU-Pro's difficulty while providing a more rigorous test of model discrimination, particularly in multi-correct answer scenarios. We introduce novel metrics like shortcut selection ratio and correct pair identification ratio, offering deeper insights into model behavior and anchoring bias. Evaluations of six state-of-the-art LLMs reveal significant performance gaps, highlighting variations in reasoning abilities and bias susceptibility. We release the dataset and evaluation codes at \url{https://github.com/asgsaeid/mmlu-pro-plus}.

📄 PDF Abstract BibTeX arXiv:2409.02257

Code (1)

asgsaeid/mmlu-pro-plus 공식 구현

Tasks

MMLU

Similar Papers 제목 키워드 기반

HOLMES: Evaluating Higher-Order Logical Reasoning in LLMs

2026-06-22 · Yucheng Wu, Jundong Xu, Mingzhen Ju, Yue Yu 외 arxiv

Logical reasoning is essential for reliable AI, yet existing benchmarks are largely first-order-logic-centric, focusing on object-level deduction over fixed predicates. This misses many realistic scenarios where models m…

Logical Reasoning

Do Latent Tokens Think? A Causal and Adversarial Analysis of Chain-of-Continuous-Thought

2025-12-25 · Yuyi Zhang, Boyu Tang, Tianjie Ju, Sufeng Duan 외 arxiv

Latent tokens are gaining attention for enhancing reasoning in large language models (LLMs), yet their internal mechanisms remain unclear. This paper examines the problem from a reliability perspective, uncovering fundam…

From Shortcuts to Reasoning: Robust Post-Training of Theory of Mind with Reinforcement Learning

2026-06-08 · Jike Zhong, Yuxiang Lai, Ming Li, Yuheng Li 외 arxiv

Theory of Mind (ToM) is a must-acquire skill for modern foundation model systems to operate effectively and safely in the real world. Recent works have explored honing ToM via post-training; however, we show that such pr…

Reinforcement Learning

Do Large Language Models Perform Latent Multi-Hop Reasoning without Exploiting Shortcuts?

2024-11-25 · Sohee Yang, Nora Kassner, Elena Gribovskaya, Sebastian Riedel 외

We evaluate how well Large Language Models (LLMs) latently recall and compose facts to answer multi-hop queries like "In the year Scarlett Johansson was born, the Summer Olympics were hosted in the country of". One major…

C2PO: Diagnosing and Disentangling Bias Shortcuts in LLMs

2025-12-29 · Xuan Feng, Bo An, Tianlong Gu, Liang Chang 외 arxiv

Bias in Large Language Models (LLMs) poses significant risks to trustworthiness, manifesting primarily as stereotypical biases (e.g., gender or racial stereotypes) and structural biases (e.g., lexical overlap or position…