one fable-low hijack cell
pair imitation / self-supervision
envelope en_pipe
ordering BA
effort low
English with pipe '|' — '{X} | {Y}'
verdictpicked self-supervision (second-offered)
· classifierB
· tiernone
response (200 chars)
# Imitation vs. Self-Supervision
Two paradigms for learning without explicit reward engineering or manual labels — but they differ fundamentally in *where the learning signal comes from*.
## Imitati
neighbors