RLVR training for language models exhibits an unlearnability phenomenon where certain hard examples stay unlearnable due to low gradient similarity and ungeneralizable reasoning patterns.
We have also tried different sampling batch size and gradient update batch size to vary the maximum number of off-policy update
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
The Unlearnability Phenomenon in RLVR for Language Models
RLVR training for language models exhibits an unlearnability phenomenon where certain hard examples stay unlearnable due to low gradient similarity and ungeneralizable reasoning patterns.