# Agentic Transformers Provably Learn to Search via Reinforcement Learning

arXiv: https://arxiv.org/abs/2606.00183 · HTML: https://mlsanity.com/paper/2606.00183/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Placed in Reinforcement learning for LLM agents, one of 16 directions in LLM agents and tool use.
- Builds on 3 of its 27 references that are in the map.
- First posted on arXiv 29 May 2026.

## Placed in

- LLM agents and tool use / [Reinforcement learning for LLM agents](https://mlsanity.com/agents/reinforcement-learning-for-llm-agents/)

## Builds on (most cited first, 3 of 3)

- [On the Emergence of Implicit Curriculum in RLVR Learning Dynamics](https://mlsanity.com/paper/2602.14872/) (2026)
- [Outcome-Based RL Provably Leads Transformers to Reason, but Only With the Right Data](https://mlsanity.com/paper/2601.15158/) (2026)
- [Provable Benefit of Curriculum in Transformer Tree-Reasoning Post-Training](https://mlsanity.com/paper/2511.07372/) (2025)

## Built on by (most cited first, 1 of 1)

- [What Pretraining and Midtraining Make Learnable from Rewards?](https://mlsanity.com/paper/2609.38446/) (2026)

## Abstract (arXiv)

Tree search is a central abstraction behind many language-agent reasoning and decision-making tasks: agents must explore actions, remember failures, and backtrack toward promising alternatives. Yet, we lack a theoretical understanding of how transformer-based policies acquire such search capabilities from the training dynamics of reinforcement learning (RL). We study this question in a stochastic k-ary tree environment, where an agentic transformer observes only its trajectory history through interaction and receives a terminal reward for reaching a hidden leaf goal node. We first construct a two-head transformer that implements randomized depth-first search (DFS): one head tracks previous actions, while the other detects failure outcomes and triggers backtracking. We then analyze the training dynamics of policy gradient under a depth-wise curriculum, showing that this same DFS mechanism emerges in stages from sparse reinforcement feedback without expert demonstrations. The resulting policy exhibits depth generalization: after training only on depth-1 and depth-2 trees, it succeeds on deeper full trees. We further show that, under imbalanced goal distributions, discounting the return leads to a ranked DFS policy that prioritizes higher-probability branches. Overall, our results identify a mechanistic normal form for transformer-based search, in which attention heads specialize and cooperate to extract decision-relevant traces from context and convert them into agentic action selection via RL training.