
OpenAI ライブストリーム
OpenAIがライブストリーム配信イベントを開催します。放送中に具体的な発表、新製品発表、またはデモンストレーションが明かされる予定です。
OpenAIが前回予告なしのライブストリームをやった時、GPT-4 Turboをドロップして、一晩で価格設定を完全に変えました

I am trying to optimize Qwen 3.6 35B A3B sampling parameters but I am having a hard time figuring out a good benchmark to do it. As to why I believe that the recommended settings may not be optimal? One reason is that they recommend the same ones for Qwen 3.5 and 3.6 yet when I upgraded to 3.6 with everything else being identical (even the same quant) 3.6 was getting stuck in tool call loops in some programmed daily tasks in which 3.5 was not and the solution was bumping the temperature up. Ano