LLM νμ μκ° μμ½ νκ°μμ 보μ΄λ μ€λ²λ© νΈν₯(overlap bias) λΆμ. LLM νμ μκ° κΈΈμ΄, μμ λ±μ νΈν₯μ κ°μ§λ©° μ λμ μ
λ ₯μ μ·¨μ½ν λ¬Έμ λ₯Ό μΈλ°νκ² λΆμνλ€.
5 sources
1 new posts
π AI λͺ¨λΈ λ²€μΉλ§ν¬
π₯οΈ Terminal-Bench 2.0 (Top 5)
π Chatbot Arena ELO (Top 5)
Source: Chatbot Arena
π§ ARC-AGI-2 λ¬μ±λ₯
π€ 84.6% β Gemini 3 Deep Think (Google)
π§ Human Panel = 100% κΈ°μ€
Source: ARC Prize Leaderboard
μ΄ 1건