Back to evaluations
Public evaluation

Sentiment Consistency Check

Moderation accuracy audit

Evaluation type
task based
Challenge
Automated Content Moderation Workflow with OpenAI Agents SDK
Difficulty
Intermediate
Rigor
Unspecified
Open linked challenge

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

Sentiment Consistency Check

Evaluates if the moderation team correctly identifies tone using Hume AI

Input format

Voice recording clip

Output format

Boolean flag and explanation