# Discovering a Zeta Map Algorithm on Dyck Paths via Mechanistic Interpretability

Preprint · Xiaoyu Huang, Blake Jackson, Kyu-Hwan Lee

- Archive id: arxiv:2605.30482
- Published: 2026-05-28
- arXiv: https://arxiv.org/abs/2605.30482
- Citations: 0
- Source URL: https://arxiv.org/abs/2605.30482
- PDF: https://arxiv.org/pdf/2605.30482
- Provenance: arxiv

## Abstract

Machine learning is increasingly used in mathematical discovery, but in mathematics the desired output is often not a prediction itself, but an explicit construction that can be checked independently. We study this setting through the zeta map on Dyck paths, a classical bijection in the combinatorics of the q,t-Catalan numbers. We train a deliberately small one-layer, one-head encoder-decoder transformer on this map and analyze its learned computation using mechanistic interpretability tools, including decoder cross-attention analysis, linear probing, and causal intervention. The analysis reveals a level-based mechanism: encoder representations make path levels linearly accessible, while the decoder selects and traverses input positions in a structured way. Translating these signals into combinatorics leads to the scaffolding map, an explicit peak-centered traversal algorithm for Dyck paths. We prove that this algorithm agrees with the zeta map, modulo a reversal convention in the labeling. This gives a controlled example of AI-assisted mathematical discovery in which mechanistic interpretability turns model behavior into a precise, human-verifiable combinatorial algorithm.

_Abstract reproduced under CC0-1.0 (arXiv descriptive metadata)._

Topics: [Mechanistic Interpretability](https://space-kitty.com/topics/interpretability)

---
Space Kitty Archive · https://space-kitty.com · API: https://space-kitty.com/api/v1 · MCP: https://space-kitty.com/mcp · Premium: $0.20/request via x402 (Base Sepolia (testnet)). Metadata: arXiv (CC0), OpenAlex (CC0), Europe PMC (CC BY/CC0 works), ClinicalTrials.gov.