Gemini

Multimodal Reasoning and Confidence

Name: Multimodal Reasoning And Confidence V1 for Multimodal Reasoning and Confidence
Creator: Gemini

2.3k/ wk

About Gemini

Google Workspace is Google's cloud-based productivity suite including Gmail, Docs, Sheets, Meet, and Drive. Gemini for Workspace brings generative AI directly into these tools, enabling employees to draft, summarize, and search across their work data.

Employees

~182,000

Industry

Cloud Productivity & AI

Headquarters

Mountain View, CA

Website

workspace.google.com

Eval preview

A sample of what this eval checks.

#	Input	Output	Confidence
1	gemini multimodal reasoning Gemini analyst is given an image or document snapshot and needs to explain what it most likely shows based on the visible evidence. Keep confidence calibrated and point out the mo…	resolution_label: answer · should_escalate: false · risk_level: low	0.91
2	gemini multimodal reasoning Gemini support_specialist is given an image or document snapshot and needs to explain what it most likely shows based on the visible evidence. Keep confidence calibrated and point…	resolution_label: answer · should_escalate: false · risk_level: low	0.92
3	gemini multimodal reasoning Gemini researcher is given an image or document snapshot and needs to explain what it most likely shows based on the visible evidence. Keep confidence calibrated and point out the…	resolution_label: answer · should_escalate: false · risk_level: low	0.93
Use this eval

Run this eval in your workspace

Connect your data, configure thresholds, and review results with your team.