Author

Karim Bouyarmane

2 papers indexed here

We haven’t gathered this author’s papers yet. Follow them and we’ll fetch their work.

Not the right person? Other researchers publish under this name.

Preprint Aug 2026

Trie Automata for Constrained Decoding over Large Finite Sets

The trie automaton is introduced, a specialized mechanism that exploits finite-set structure via Aho-Corasick multi-pattern matching to precompute per-node token masks, and achieves 7X faster per-step valid-token computation compared to XGrammar and 2.5X faster compilation at K>= 300.

Xingzi Xu, Karim Bouyarmane · 0 citations

#artificial intelligence Preprint Sep 2026

Extremely Sparse Supervision Incentivizes Reasoning Ability

Large language models demonstrate increasingly strong reasoning capabilities through effective post-training. Yet, prevailing post-training methods optimize over massive numbers of tokens, implicitly assuming that effective learning must be token-intensive. We revisit this assumption in the on-policy distillation (OPD)...

Zhi-Shuai Liu, Xingzi Xu, Mehmet Saygin Seyfioglu et al. · 2 citations