Back to evaluations
Public evaluation

nuclear_verdict_probability

Evaluates probability score calibration for high-severity legal nuclear verdicts.

Evaluation type
task based
Challenge
Legal Nuclear Verdict Predictor using LangChain and LlamaIndex
Difficulty
Advanced
Rigor
Unspecified

Evaluation overview

How the linked challenge is judged: tasks, benchmarks, and criteria count.

Tasks
1
Benchmarks
0
Criteria
0

Task templates

Inputs and expected outputs.

Task 1

nuclear_verdict_probability

Predicts nuclear verdict likelihood for a casualty claim file.

Input format

JSON containing claim_summary, jurisdiction, plaintiff_demands_m, and injury_severity_code

Output format

JSON with nuclear_verdict_prob (float 0-1), primary_risk_jurisdiction (string), and expected_severity_tier (string)