Back to evaluations
Draft evaluation
AutoGen Casualty ILS: Multi-Agent Tail Risk Structuring System — evaluation
Evaluates multi-agent conversation termination and output yield pricing error against ground truth actuary benchmarks.
Evaluation type
task based
Challenge
AutoGen Casualty ILS: Multi-Agent Tail Risk Structuring System
Difficulty
Advanced
Rigor
Not declared
The author has not specified a rigor level.
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
casualty_ils_structuring
Agents negotiate ILS bond parameters and output agreed coupon spread.
Input format
JSON containing expected_loss_pct, attachment_point, tail_factor
Output format
JSON containing agreed_coupon_bps, rating_grade, market_clearing_status