Author

Sanghyeok Park

3 papers indexed here

We haven’t gathered this author’s papers yet. Follow them and we’ll fetch their work.

Not the right person? Other researchers publish under this name.

#natural language process... Preprint Sep 2026

SWORD: Wikidata-based Distortions Reveal Hidden Cross-Lingual Inconsistencies in LLM Factual Error Rejection

This work introduces Systematic Wikidata-based Object-Relation Distortion (SWORD), a benchmark that evaluates whether models consistently reject factual errors across languages, and demonstrates that multilingual factual reasoning involves asymmetric capabilities that aggregate accuracy metrics systematically obscure.

Sanghyeok Park, Min-Ji Kang, Hosung Kwak et al. · 0 citations

#artificial intelligence Jun 2026

Rollout-Level Advantage-Prioritized Experience Replay for GRPO

A rollout-level replay buffer for GRPO is proposed that stores and samples individual rollouts rather than whole groups and bounds staleness through age eviction, and is the only condition with a positive margin over GRPO at every scale.

Gyeongtae Yoo, Sanghyeok Park, Soohyuk Jang et al. · arXiv.org · 2 citations

#artificial intelligence Preprint Aug 2026

SHADOWBENCH: Toward Reliable Automatic Evaluation of Semantic Alignment in Autoformalization

SA-Pass (*Semantic Alignment Pass*), which tests formal statements using auxiliary statements called shadows that characterize the intended statement, is proposed, which achieves binary agreement with expert judgments.

Hojae Han, Jongyoon Kim, Sanghyeok Park et al. · 2 citations