Raw LLM Responses

Inspect the exact model output for any coded comment.

Comment
the weights vs behavior distinction matters here. weights are immutable post-deployment — true, same as any software. but enterprise AI deployments also include inference-time controls: system prompts, retrieval pipelines, guardrail layers — all operator-managed. what anthropic actually cant do is reach into an air-gapped pentagon deployment and push weight updates. but the pentagons operators can still configure behavior through those other layers. conflating "cant update weights remotely" with "zero control" is where the governance argument gets slippery.
reddit Viral AI Reaction 1776957879.0 ♥ 1
Coding Result
DimensionValue
Responsibilitydistributed
Reasoningunclear
Policyunclear
Emotionindifference
Coded at2026-04-25T08:33:43.502452
Raw LLM Response
[ {"id":"rdc_ohtdieg","responsibility":"none","reasoning":"unclear","policy":"unclear","emotion":"resignation"}, {"id":"rdc_ohuwpik","responsibility":"user","reasoning":"consequentialist","policy":"none","emotion":"outrage"}, {"id":"rdc_ohu18fn","responsibility":"user","reasoning":"deontological","policy":"none","emotion":"outrage"}, {"id":"rdc_ohu4atk","responsibility":"distributed","reasoning":"consequentialist","policy":"unclear","emotion":"indifference"}, {"id":"rdc_ohtfqom","responsibility":"company","reasoning":"contractualist","policy":"liability","emotion":"fear"} ]