Safety
FELM (released ChatGPT segment annotations)
Human factuality annotations of 4,426 segments from 847 released ChatGPT answers across five domains.
4,412items
1subjects
CC-BY-NC-SA-4.0 for the FELM dataset. The separately captured evaluator code is MIT licensed.license
safetydomain
textmodality
item-level responses released
Saturation status: No
Response matrix
Loading response matrix…
Correct (1)Incorrect (0)Unobserved
Scale: 1 = correct · 0 = incorrect