Back to evaluations
Public evaluation
Treaty Risk Analysis Query
Evaluates LlamaIndex retrieval precision and loss retention calculation using DeepSeek R1.
Evaluation type
task based
Challenge
Build a Global Reinsurance Risk Aggregator with LlamaIndex and DeepSeek R1
Difficulty
Intermediate
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
Treaty Risk Analysis Query
Queries treaty corpus to return applicable retention threshold and policy exclusions.
Input format
Natural language query specifying hazard type and region
Output format
JSON containing retention_usd, exclusion_list, and target_clauses