Raw LLM Responses

Inspect the exact model output for any coded comment.

Comment
It's hard to tell but I think I noticed the same thing this week, unprompted by any Reddit posts or anything. It just seems more reluctant to talk on anything that was even slightly controversial or negative, whereas before I could swear it was functioning smoothly. I'm also 95% sure this happened with gpt3, I think anybody who remembers what gpt3 was like at lunch would back this.
reddit AI Harm Incident 1681482582.0 ♥ 4
Coding Result
DimensionValue
Responsibilitycompany
Reasoningconsequentialist
Policyunclear
Emotionmixed
Coded at2026-04-25T08:33:43.502452
Raw LLM Response
[ {"id":"rdc_jg75s2w","responsibility":"developer","reasoning":"consequentialist","policy":"none","emotion":"resignation"}, {"id":"rdc_jg7appc","responsibility":"unclear","reasoning":"unclear","policy":"unclear","emotion":"indifference"}, {"id":"rdc_jg8e8i4","responsibility":"company","reasoning":"consequentialist","policy":"unclear","emotion":"mixed"}, {"id":"rdc_jg7qdbq","responsibility":"unclear","reasoning":"unclear","policy":"unclear","emotion":"approval"}, {"id":"rdc_jg81vbb","responsibility":"company","reasoning":"deontological","policy":"none","emotion":"outrage"} ]