MIT 最新研究發現，大型語言模型（LLM）回答問題時有時會「學錯重點」，依賴訓練期間學到的語法模式作答，而非真正理解問題內容。這種現象會令模型處理新任務時出現意外失誤，影響客戶查詢處理、臨床記錄摘要及財務報告
據內媒報道，新加坡國家人工智能計劃（AISG）正進行一次重大戰略調整，在其最新東南亞語言大模型項目中放棄Meta模型，轉向阿里（9988）旗下通義千問（Qwen）開源架構，標誌着中國開源AI模型在全球影響力版圖中的一次關
Anthropic has announced its latest AI model with Claude Opus 4.5. The company has also expanded Claude Code availability to the Claude desktop app for the first time.
Anthropic describes its new Opus 4.5 model as “intelligent, efficient, and the best model in the world for coding, agents, and computer use.” It follows Opus 4.1, which Anthropic released in August. The company shares internal first impressions of new model:
As our Anthropic colleagues tested the model before release, we heard remarkably
Apple researchers have published a study that looks into how LLMs can analyze audio and motion data to get a better overview of the user’s activities. Here are the details.
A new paper titled “ ” offers insight into how Apple may be considering incorporating LLM analysis alongside traditional sensor data to gain a more precise understanding of user activity.
This, they argue, has great potential to make activity analysis more precise, even in situations where there isn’t enough sensor data.
From the researchers:
“Sensor data streams