Raw LLM Responses

Inspect the exact model output for any coded comment.

Comment
It's on the verge of no longer being useful and we are debating canceling our work subscriptions (several hundred bucks a month plus APIs). A few months back, even the free GPT-3.5 model produced better results. Questions regarding software development, one of ChatGPTs known strong suits, often result in hallucinations. It invents things, and when asked about it, apologizes. Mere weeks ago, results were very stable and precise. Yesterday, I was brainstorming for a legal document to work on with my lawyer. Instead of helping me come up with ideas, every other sentence would be followed with a canned response that I should talk to a lawyer. No shit boy...
reddit AI Harm Incident 1689786421.0 ♥ 5
Coding Result
DimensionValue
Responsibilitycompany
Reasoningconsequentialist
Policynone
Emotionunclear
Coded at2026-04-25T08:33:43.502452
Raw LLM Response
[{"id":"rdc_jslc385","responsibility":"ai_itself","reasoning":"consequentialist","policy":"none","emotion":"indifference"},{"id":"rdc_jslo95t","responsibility":"none","reasoning":"deontological","policy":"none","emotion":"outrage"},{"id":"rdc_jsm370w","responsibility":"company","reasoning":"consequentialist","policy":"none","emotion":"frustration"},{"id":"rdc_jsmea4a","responsibility":"company","reasoning":"virtue","policy":"none","emotion":"outrage"},{"id":"rdc_jsk7i6o","responsibility":"none","reasoning":"consequentialist","policy":"none","emotion":"indifference"}]