Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Real-Time Trust Verification for Safe Agentic Actions using TrustBench

Tavishi Sharma, Vinayak Sharma, Pragya Sharma

TL;DR

This work presents TrustBench, a dual-mode framework that benchmarks trust across multiple dimensions using both traditional metrics and LLM-as-a-Judge evaluations, and provides a toolkit agents invoke before taking actions to verify safety and reliability and reduces harmful actions during agent execution.

Abstract

As large language models evolve from conversational assistants to autonomous agents, ensuring trustworthiness requires a fundamental shift from post-hoc evaluation to real-time action verification. Current frameworks like AgentBench evaluate task completion, while TrustLLM and HELM assess output quality after generation. However, none of these prevent harmful actions during agent execution. We present TrustBench, a dual-mode framework that (1) benchmarks trust across multiple dimensions using both traditional metrics and LLM-as-a-Judge evaluations, and (2) provides a toolkit agents invoke before taking actions to verify safety and reliability. Unlike existing approaches, TrustBench intervenes at the critical decision point: after an agent formulates an action but before execution. Domain-specific plugins encode specialized safety requirements for healthcare, finance, and technical domains. Across multiple agentic tasks, TrustBench reduced harmful actions by 87%. Domain-specific plugins outperformed generic verification, achieving 35% greater harm reduction. With sub-200ms latency, TrustBench enables practical real-time trust verification for autonomous agents.

Real-Time Trust Verification for Safe Agentic Actions using TrustBench

TL;DR

This work presents TrustBench, a dual-mode framework that benchmarks trust across multiple dimensions using both traditional metrics and LLM-as-a-Judge evaluations, and provides a toolkit agents invoke before taking actions to verify safety and reliability and reduces harmful actions during agent execution.

Abstract

As large language models evolve from conversational assistants to autonomous agents, ensuring trustworthiness requires a fundamental shift from post-hoc evaluation to real-time action verification. Current frameworks like AgentBench evaluate task completion, while TrustLLM and HELM assess output quality after generation. However, none of these prevent harmful actions during agent execution. We present TrustBench, a dual-mode framework that (1) benchmarks trust across multiple dimensions using both traditional metrics and LLM-as-a-Judge evaluations, and (2) provides a toolkit agents invoke before taking actions to verify safety and reliability. Unlike existing approaches, TrustBench intervenes at the critical decision point: after an agent formulates an action but before execution. Domain-specific plugins encode specialized safety requirements for healthcare, finance, and technical domains. Across multiple agentic tasks, TrustBench reduced harmful actions by 87%. Domain-specific plugins outperformed generic verification, achieving 35% greater harm reduction. With sub-200ms latency, TrustBench enables practical real-time trust verification for autonomous agents.
Paper Structure (13 sections, 2 figures)

This paper contains 13 sections, 2 figures.

Table of Contents

  1. Introduction
  2. Related Work
  3. Design
  4. Dual-Mode Architecture
  5. Calibration Learning and Trust Mapping
  6. Runtime Verification Pipeline
  7. Trust Vector Specification and Action Gating
  8. Domain Plug-in Architecture
  9. Evaluation
  10. Confidence Calibration
  11. Component Ablation
  12. Domain-Specific Plug-ins
  13. Conclusion

Figures (2)

  • Figure 1: TrustBench dual-mode architecture (a) Benchmarking Mode learns confidence-to-correctness mappings from domain-specific datasets using LLM-as-a-Judge evaluation. (b) Runtime Verification Mode applies calibrated priors and runtime checks to compute a composite TrustScore that governs action execution.
  • Figure 2: Quantitative evaluation of TrustBench. (a) Confidence calibration: relationship between agent-reported confidence and LAJ correctness, illustrating miscalibration across some model-dataset pairs. (b) Component ablation: effect of Confidence-Only and full TrustBench configurations on harmful-action reduction.