Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Real-Time Trustworthiness Scoring for LLM Structured Outputs and Data Extraction

Hui Wen Goh, Jonas Mueller

Abstract

Structured Outputs from current LLMs exhibit sporadic errors, hindering enterprise AI efforts from realizing their immense potential. We present CONSTRUCT, a method to score the trustworthiness of LLM Structured Outputs in real-time, such that lower-scoring outputs are more likely to contain errors. This reveals the best places to focus limited human review bandwidth. CONSTRUCT additionally scores the trustworthiness of each field within a LLM Structured Output, helping reviewers quickly identify which parts of the output are wrong. Our method is suitable for any LLM (including black-box LLM APIs without logprobs such as reasoning models and Anthropic models), does not require labeled training data nor custom model deployment, and works for complex Structured Outputs with many fields of diverse types (including nested JSON schemas). We additionally present one of the first public LLM Structured Output benchmarks with reliable ground-truth values that are not full of mistakes. Over this four-dataset benchmark, CONSTRUCT detects errors from various LLMs (including Gemini 3 and GPT-5) with significantly higher precision/recall than other scoring methods.

Real-Time Trustworthiness Scoring for LLM Structured Outputs and Data Extraction

Abstract

Structured Outputs from current LLMs exhibit sporadic errors, hindering enterprise AI efforts from realizing their immense potential. We present CONSTRUCT, a method to score the trustworthiness of LLM Structured Outputs in real-time, such that lower-scoring outputs are more likely to contain errors. This reveals the best places to focus limited human review bandwidth. CONSTRUCT additionally scores the trustworthiness of each field within a LLM Structured Output, helping reviewers quickly identify which parts of the output are wrong. Our method is suitable for any LLM (including black-box LLM APIs without logprobs such as reasoning models and Anthropic models), does not require labeled training data nor custom model deployment, and works for complex Structured Outputs with many fields of diverse types (including nested JSON schemas). We additionally present one of the first public LLM Structured Output benchmarks with reliable ground-truth values that are not full of mistakes. Over this four-dataset benchmark, CONSTRUCT detects errors from various LLMs (including Gemini 3 and GPT-5) with significantly higher precision/recall than other scoring methods.
Paper Structure (60 sections, 2 equations, 6 figures, 1 table)

This paper contains 60 sections, 2 equations, 6 figures, 1 table.

Table of Contents

  1. Introduction
  2. Related Work
  3. Scoring the Trustworthiness of Structured Outputs from any LLM
  4. Principles for an effective trustworthiness score
  5. How CONSTRUCT works
  6. A High-Quality Structured Outputs Benchmark
  7. Dataset: Financial Entities Extraction
  8. Dataset: PII Extraction
  9. Dataset: Insurance Claims Extraction
  10. Dataset: Data Table Analysis
  11. Benchmarking LLM Accuracy
  12. Benchmarking Trustworthiness Scores
  13. Setup
  14. Results
  15. Discussion
  16. ...and 45 more sections

Figures (6)

  • Figure 1: How effectively do different Per-Document scoring techniques detect incorrect LLM Structured Outputs. Here a document's LLM output is deemed incorrect if any of its fields are wrong, the subtitle of each graph indicates which model produced the outputs, and detection effectiveness is measured via AUROC (higher is better).
  • Figure 2: How effectively (in terms of AUROC) do different Per-Field scoring techniques detect individual fields that are erroneous within LLM Structured Outputs.
  • Figure 3: Precision @ Num-Errors achieved by various Per-Document scores applied to outputs generated from different models. Here higher values are better.
  • Figure 4: Precision @ Num-Errors achieved by various Per-Field scores applied to outputs generated from different models. Here higher values are better.
  • Figure 5: Confidence Gap achieved by various Per-Document scores applied to outputs generated from different models. Here higher values are better.
  • ...and 1 more figures