# RadixArk praises Sasha's Lean-verified Rust tokenizer for agentic RL

> Original title: Love seeing ideas from Miles' Token-In-Token-Out work inspire new research!

- Source: RadixArk
- Published: 2026-10-28T00:04:30.000Z
- Bad Influence: https://badinfluence.news/items/ft0tjmrndkjf1wep7vdudpjss
- Original: https://x.com/radixark/status/2115233208438776259

## Summary

RadixArk praises Sasha's work formally verifying properties of a Rust tokenizer in Lean, which extends Miles' token-in-token-out idea beyond testing that training sees the exact tokens the model generated. Sasha's prototype, posted by Sasha Rush (@srush_nlp), addresses training-inference mismatch in RL by verifying the tokenizer pipeline rather than relying on messages built in Jinja.

## Post

Love seeing ideas from Miles' Token-In-Token-Out work inspire new research!

Instead of just testing that training sees the exact tokens the model generated, Sasha takes it a step further, formally verifying key properties of a Rust tokenizer in Lean. That's the kind of rigor agentic RL needs.

Fascinating work, @srush_nlp!

## Quoted post @srush_nlp

> Lean Verified Tokenizers (https://srush.github.io/verified-tokenizer/)
> 
> Training-inference mismatch kills RL runs, but yet we build messages in Jinja? Prototypes a Lean-verified Rust agent pipeline.

https://x.com/srush_nlp/status/2115197402525434242