This archive copy may be incomplete. The original-source link is preserved in the card footer.
RT @Mascobot: Even if your LLM can do 128K context window, you still want to keep the context as short as possible. The performance drop…
More notes on nearby ideas
Tweet · Mar 4, 2025do your very long context llm chats tend to get...Quote tweet · Jun 15, 2026lol, i feel like there's a whole cluster of weird behavior downstream of the gap between how humans experience time vs how llms experience the context…Artificial intelligenceRepost · Mar 30, 2023RT @KateandPie: @karpathy - You dont have a short attention span, you have a limited context windowRepost · Dec 4, 2022RT @garybasin: Current arch are memory-constrained — maybe a few thousand tokens of context. Assuming this remains binding, how do AIs adap…