TLPO mitigates language confusion in LLMs via token-level policy updates that outperform sequence-level methods while preserving general capabilities.
InSecond Conference on Language Modeling
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
TLPO: Token-Level Policy Optimization for Mitigating Language Confusion in Large Language Models
TLPO mitigates language confusion in LLMs via token-level policy updates that outperform sequence-level methods while preserving general capabilities.