Event JSON
{
"id": "5792aaacf1ee29bbe77790a773ad93436276261172f4633360ca8c05d8be41ae",
"pubkey": "e8b038ee9a4a2ca0aece0fdd06407ef435a38aad8d2597bb32f63f0cbe01019e",
"created_at": 1746164727,
"kind": 1,
"tags": [
[
"t",
"llm"
],
[
"imeta",
"url https://i.maly.io/media_attachments/files/114/436/651/476/307/508/original/ba3ad931e20e254c.webp",
"m image/webp",
"dim 1024x623",
"blurhash U5R{#;%NM{?b_4aeogD%Dhxa-=IV_3WDoM-;"
],
[
"proxy",
"https://maly.io/users/MattiSG/statuses/114436651591448807",
"activitypub"
],
[
"client",
"Mostr",
"31990:6be38f8c63df7dbf84db7ec4a6e6fbbd8d19dca3b980efad18585c46f04b26f9:mostr",
"wss://relay.mostr.pub"
]
],
"content": "#LLM “refusal to answer is the default behavior: we find a circuit that is on by default and that causes the model to state that it has insufficient information to answer any given question. However, when the model is asked about something it knows well, a competing feature representing ‘known entities’ activates and inhibits this default circuit”\nhttps://www.anthropic.com/research/tracing-thoughts-language-model\n\nhttps://i.maly.io/media_attachments/files/114/436/651/476/307/508/original/ba3ad931e20e254c.webp",
"sig": "02d5ddcdd0308ec999963852015bf8989d97efae6959364926e845185ff15be33d29bab0ebd9333cc640a52b9fa7fe3133c8fcd5d733cd82d9385159c89eb623"
}