Why Does Self-Distillation (Sometimes) Degrade the Reasoning Capability of LLMs?
Paper • 2603.24472 • Published • 47
Jeongjae Park
jjp97
AI & ML interests
I’m interested in the latest NLP and AI technologies, such as uncertainty, retrieval, agentic approaches, and long-context models!
Recent Activity
upvoted a paper about 1 hour ago
Why Does Self-Distillation (Sometimes) Degrade the Reasoning Capability of LLMs? upvoted a paper 10 days ago
Efficient Reasoning with Balanced Thinking upvoted a paper 10 days ago
Attention ResidualsOrganizations
upvoted a paper about 1 hour ago
upvoted 2
papers 10 days ago
upvoted a paper 14 days ago
liked a
model 17 days ago
upvoted 2
papers 20 days ago
upvoted 3
papers about 1 month ago
liked 5
models about 2 months ago
allenai/OLMoE-1B-7B-0125
Text Generation • 7B • Updated • 29.4k • 35
deepseek-ai/deepseek-moe-16b-base
Text Generation • 16B • Updated • 23.9k • 142
Qwen/Qwen1.5-MoE-A2.7B
Text Generation • 14B • Updated • 199k • 222
microsoft/Phi-3.5-MoE-instruct
Text Generation • Updated • 94.1k • 571
allenai/OLMoE-1B-7B-0125-Instruct
Text Generation • 7B • Updated • 52.2k • 61
upvoted 5
papers about 2 months ago
Shaping capabilities with token-level data filtering
Paper • 2601.21571 • Published • 27
A-RAG: Scaling Agentic Retrieval-Augmented Generation via Hierarchical Retrieval Interfaces
Paper • 2602.03442 • Published • 21
PaperBanana: Automating Academic Illustration for AI Scientists
Paper • 2601.23265 • Published • 222
Why Language Models Hallucinate
Paper • 2509.04664 • Published • 199
Native Parallel Reasoner: Reasoning in Parallelism via Self-Distilled Reinforcement Learning
Paper • 2512.07461 • Published • 79