Raw LLM Responses

Inspect the exact model output for any coded comment.

Comment
As for Claude 4 having an "instinct for self-preservation" I'm not so sure that is an accurate conclusion. This is just my amateur's guess but it seems more likely to me that the AI might have learned to copy such a response in it's data about human interactions and behaviors. In other words humans are possibly creating a bad example to AIs for how to respond to certain situations.
youtube AI Governance 2025-08-26T17:1…
Coding Result
DimensionValue
Responsibilitydeveloper
Reasoningmixed
Policyunclear
Emotionmixed
Coded at2026-04-26T19:39:26.816318
Raw LLM Response
[ {"id":"ytc_Ugz3qTS819wIgZshvBl4AaABAg","responsibility":"government","reasoning":"consequentialist","policy":"regulate","emotion":"fear"}, {"id":"ytc_UgyWwj-vUslVBQdFn354AaABAg","responsibility":"developer","reasoning":"deontological","policy":"liability","emotion":"outrage"}, {"id":"ytc_Ugz4anTgjdsGbSssSZJ4AaABAg","responsibility":"developer","reasoning":"mixed","policy":"unclear","emotion":"mixed"}, {"id":"ytc_UgwaHX4mvwUBpBLGR8J4AaABAg","responsibility":"ai_itself","reasoning":"consequentialist","policy":"unclear","emotion":"fear"}, {"id":"ytc_UgzIpceCfSqOdxPm3mx4AaABAg","responsibility":"government","reasoning":"consequentialist","policy":"ban","emotion":"fear"} ]