Raw LLM Responses

Inspect the exact model output for any coded comment.

Comment
"We found instances of the model attempting to write self-propagating worms, fabricating legal documentation, and leaving hidden notes to future instances of itself all in an effort to undermine its developers' intentions," Apollo Research said in notes included as part of Anthropic's safety report for Opus 4. * This is a quote from the Axios article regarding the Claude 4 Opus model Anthropic just released. After testing, Apollo Research told them NOT to release the model externally OR internally. The CEO is sounding the alarm because he knows what’s about to happen is out of their control. I believe he’s way past being concerned. Deep down he’s probably terrified. And Anderson Cooper didn’t press him harder about the article. Shocker.
youtube AI Jobs 2025-06-02T20:1…
Coding Result
DimensionValue
Responsibilitydeveloper
Reasoningunclear
Policyunclear
Emotionfear
Coded at2026-04-26T19:39:26.816318
Raw LLM Response
[ {"id":"ytc_UgxwD2T72hKsyLvuvRp4AaABAg","responsibility":"none","reasoning":"consequentialist","policy":"none","emotion":"approval"}, {"id":"ytc_UgxT5C7yidaLWLorh-94AaABAg","responsibility":"developer","reasoning":"unclear","policy":"unclear","emotion":"fear"}, {"id":"ytc_UgxiuQh-7R2yRr6wpxh4AaABAg","responsibility":"user","reasoning":"virtue","policy":"none","emotion":"indifference"}, {"id":"ytc_UgxLun61AgZ218upkFJ4AaABAg","responsibility":"company","reasoning":"consequentialist","policy":"none","emotion":"outrage"}, {"id":"ytc_UgwuuQ6UzB7cmRX7uHR4AaABAg","responsibility":"none","reasoning":"consequentialist","policy":"none","emotion":"resignation"} ]