NAVER AI Lab · 14 records

dateidtitleauthorsvenue↑
2026-11-192611.17460Unfinished-Bench: Measuring Rule Violations by Persistent Agents in Realistic Work EnvironmentsSeunghyun Lee et al.arXiv8
2026-11-132611.16905Factored One-sided Adam-Moment for Practical and Scalable SOAPJaemyung Yu et al.arXiv14
2026-10-212610.27851LaLa-World: Scaling Latent World Models with Natural Language ActionsJunha Song et al.arXiv50
2026-10-162610.22347Speculative Forcing for Efficient Autoregressive Video GenerationHyunwook Choi et al.arXiv80
2026-10-142610.14351Learning Transferable World-Action Models from Task-Paired Human VideosJeongeun Park et al.arXiv12
2026-10-142610.16135Omni-Modal LLMs Know More Than They SayNa Min An et al.arXiv10
2026-10-082610.07874On-Policy Distillation with Negative-Policy RolloutsJaehui Hwang et al.arXiv21
2026-10-082610.07659DLoop: Looped Speculative DecodingGeonmo Gu et al.arXiv15
2026-09-102609.07103Revisiting Complete Reasoning Traces for Post-TrainingJaehui Hwang et al.arXiv20
2026-09-072609.04753Beneath the Surface of Chains-of-Thought: A Mechanistic Interpretation of Reasoning Operations in LLMsSeogyeong Jeong et al.arXiv14
2026-09-012608.30135Verification-Aware Training for Speculative DecodingGeonmo Gu et al.arXiv10
2026-08-072608.05802On-Policy Delta Distillation for Multilingual Math ReasoningByeongho Heo et al.arXiv10
2026-07-202607.15161On-Policy Delta DistillationByeongho Heo et al.arXiv35
2026-06-162606.15631Retrieve, Don't Retrain: Extending Vision Language Action Models to New Tasks at Test TimeJeongeun Park et al.arXiv16