LLMs Papers Overview
Key Themes
- Foundation Models: GPT-3, 4o, LLaMA, Gemini
- Efficient Fine-tuning: LoRA, QLoRA, LoraHub
- MoE Architecture: Mixtral, V3
- Reasoning: CoT Prompting, Reasoning Models Struggle to Control their Chains of Thought
- Evaluation: LLM-as-Judge Survey
Related Concepts
Cross-Domain Connections
- Reasoning: Reasoning Survey - CoT, faithful reasoning 연구
- Self-Evolving: Self-Evolving Survey - LLM 자기개선 메커니즘
- Agents: Agent Papers - LLM 기반 에이전트 시스템
- Self-Consciousness: Self-Consciousness - LLM의 자기인식
- RL: Reinforcement Learning - RLHF, PPO, DPO 관련
- Statistics: 통계 - 모델 평가 시 통계적 검정