paper-with-me

Papers

KCSAT-ML: Probing Reasoning Models with Nationwide-Cohort Human Difficulty

2026-06-09 · Sanghee Park, Geewook Kim, Kee-Eung Kim arxiv

Math reasoning benchmarks have proliferated, yet most lack a per-item difficulty signal grounded in actual human performance. We introduce KCSAT-ML, a decade (2014-2025) of Korean College Scholastic Ability Test (KCSAT; Suneung) mathematics: 664 problems with a 339-item core set carrying official per-item error rates from nationwide cohorts of hundreds of thousands of examinees. We pair the benchmark with Difficulty-aligned Reasoning Gain (DRG): a score-orthogonal metric that asks whether a model's mistakes concentrate on the items humans found hard, or on items humans found easy. Together they expose, across a wide range of VLMs (and LLMs with OCR), three patterns: (i) low-budget accuracy collapses on the high-human-error tail at every model size; (ii) test-time scaling (TTS) raises token use roughly linearly with cohort error rate, while accuracy gains follow a non-monotonic curve; (iii) within a single family, TTS flips between anti-scaling on the hardest items and overthinking on easier ones -- two faces of the same alignment failure. With the proposed DRG metric, we find that models with near-identical accuracy can sit at near-opposite values: one model gets wrong what humans also find hard, while another solves the hardest items yet fails on items humans find easy, which is a critical contrast that aggregate accuracy hides. Our code and dataset builder is fully open-sourced at https://github.com/naver-ai/KCSAT-ML.

📄 PDF Abstract BibTeX arXiv:2606.10403

Code (0)

등록된 구현이 없습니다.

Similar Papers 제목 키워드 기반

Enhancing LLMs' Clinical Reasoning with Real-World Data from a Nationwide Sepsis Registry

2025-05-05 · Junu Kim, Chaeeun Shim, Sungjin Park, Su Yeon Lee 외

Although large language models (LLMs) have demonstrated impressive reasoning capabilities across general domains, their effectiveness in real-world clinical practice remains limited. This is likely due to their insuffici…

Probing Commonsense Explanation in Dialogue Response Generation

2021-04-19 · Findings (EMNLP) 2021 11 · Pei Zhou, Pegah Jandaghi, Bill Yuchen Lin, Justin Cho 외

Humans use commonsense reasoning (CSR) implicitly to produce natural and coherent responses in conversations. Aiming to close the gap between current response generation (RG) models and human communication abilities, we …

Common Sense ReasoningResponse Generation

Nationwide EHR-Based Chronic Rhinosinusitis Prediction Using Demographic-Stratified Models

2026-04-16 · Sicong Chang, Yidan Shen, Justina Varghese, Akshay R Prabhakar 외 arxiv

Chronic rhinosinusitis (CRS) is a common heterogeneous inflammatory disorder that causes substantial morbidity and healthcare costs. CRS is difficult to identify early from routine encounters, as symptom presentations ov…

Machine-Assisted Grading of Nationwide School-Leaving Essay Exams with LLMs and Statistical NLP

2026-01-22 · Andres Karjus, Kais Allkivi, Silvia Maine, Katarin Leppik 외 arxiv

Large language models (LLMs) enable rapid and consistent automated evaluation of open-ended exam responses, including dimensions of content and argumentation that have traditionally required human judgment. This is parti…

Honeyquest for LLMs: Rethinking Cyber Deception for AI Attackers

2026-06-19 · Kerri Prinos, Lilianne Brush, Cameron Denton arxiv

The empirical foundation of cyber deception relies on human-centered hypotheses, but the rapid emergence of autonomous, AI-enabled attackers challenges whether this foundation transfers to AI agents. To address this, we …