TL;DR: We probed SONAR text autoencoders to see if they implicitly learn "correctness" across domains. Turns out they do, but with a clear hierarchy: code validity (96% accuracy) > grammaticality (93%
Investigating Internal Representations of…
TL;DR: We probed SONAR text autoencoders to see if they implicitly learn "correctness" across domains. Turns out they do, but with a clear hierarchy: code validity (96% accuracy) > grammaticality (93%