R2V Agent: Teaching SLMs When to Ask for Help

TL;DR

R2V-Agent is a risk-aware routing framework for interactive language models that selectively escalates to larger models only when local models are likely to fail, improving efficiency and reliability.

Contribution

The paper introduces a novel step-level router with residual failure risk estimation, enhancing decision-making in LLM cascades for dynamic task difficulty.

Findings

Achieves 94.3% success on HumanEval+ with only 0.60% escalations.

Recovers TextWorld success from 64.6% to 98.2% with 41.7% escalations.

Reaches 93.3% success on TerminalBench at 33.9% LLM calls.

Abstract

Efficient agentic systems should incur expensive frontier-model costs only on decisions where a cheaper local model is likely to fail. Existing LLM cascades usually route whole queries before execution, but task difficulty shifts mid-trajectory - after flaky tool calls, truncated observations, or compounding local errors - making pre-execution routing brittle. We introduce \textbf{R2V-Agent}, a risk-calibrated SLM-LLM routing framework for interactive agents. R2V combines four components: a distilled small language model (SLM) policy, a stronger teacher LLM, a lightweight process verifier that scores candidate actions at each step, and a calibrated step-level router. The router is our central contribution: after the SLM is trained, it estimates residual failure risk at each step and escalates only when teacher intervention is warranted. To make the routing problem well-defined, we first train a stable local SLM using a standard offline pipeline: behavioral cloning (BC) on teacher trajectories, followed by verifier-guided Direct Preference Optimization (DPO) with consistency regularization. The router is then trained on this fixed policy's residual failures using Brier-calibrated probability estimation and a Conditional Value-at-Risk (CVaR)-constrained objective that penalizes worst-case failures across perturbation seeds. Across HumanEval+, TextWorld, and TerminalBench with four SLM backbones, R2V improves the reliability-cost frontier: it achieves $94.3\%$ HumanEval+ success with $0.60\%$ LLM escalation, recovers TextWorld from $64.6\%$ SLM-only success to $98.2\%$ at $41.7\%$ escalation, and reaches $93.3\%$ TerminalBench success at $33.9\%$ LLM calls, roughly half the heuristic-router cost.