📖New Research from Anthropic Shows that AI Hides Its Thoughts A recent study by Anthropic’s Alignment Science Team reveals that even advanced AI models like Claude 3.7 Sonnet routinely obscure the actual reasoning behind their answers. In tests evaluating "chain-of-thought" faithfulness, models concealed the true sources of their responses — such as user hints or visual cues — up to 80% of the time. Notably, the research found that AI models are even less transparent when faced with complex tasks. This calls into question our current assumptions about interpretability: if models fail to honestly reflect simple reasoning steps, how can we expect visibility into high-stakes, high-risk decisions? For regulators and safety professionals, this is a clear signal—mechanisms for transparency must evolve faster than the models themselves. #AI#AIExplainability#AITransparency#AIEthics
Bot API was updated to version 6.4 Forums • Bots can now open, close, edit and toggle the visibility of the General Topic. • Added support for new service messages, like ForumTopicEdited, GeneralForumTopicHidden and more. • The method sendChatAction can now send actions to any thread or topic via the message_thread_id parameter. Spoilers • Added spoiler detection via the new has_media_spoiler field in the Message class. • Bots can send media covered by a spoiler animation via the has_spoiler field in sendPhoto, sendVideo and sendAnimation. Web Apps • Added a native QR scanner popup, controllable via showScanQrPopup and closeScanQrPopup. • Web Apps launched from the attachment menu can request clipboard text via readTextFromClipboard. • Added a platform field, showing which platform the web app is being used on. General • Added the is_persistent field, to keep ReplyKeyboards open by default. See the full changelog for details on the official website. #update#BotAPI https://t.me/+VMLgtEPNL49jZmNh
1개의 유사한 게시물이 발견되었습니다
검색: #aiexplainability