CoolFace
Modelpublic

procedure2012/Cobalt-Safety-Guard

sourceHugging Faceapache-2.0updated 2mo agoView on Hugging Face
0likes9downloads
Model Card

Cobalt-Safety-Guard

<!-- markdownlint-disable first-line-h1 --> <!-- markdownlint-disable html --> <!-- markdownlint-disable no-duplicate-header -->

<div align="center"> <img src="figures/fig1.png" width="60%" alt="Cobalt-Safety-Guard" /> </div> <hr>

<div align="center" style="line-height: 1;"> <a href="LICENSE" style="margin: 2px;"> <img alt="License" src="figures/fig2.png" style="display: inline-block; vertical-align: middle;"/> </a> </div>

1. Introduction

Cobalt-Safety-Guard is a moderation-aware model post-trained with a heavy emphasis on safety and refusal calibration.

<p align="center"> <img width="80%" src="figures/fig3.png"> </p>

2. Evaluation Results

Comprehensive Benchmark Results

<div align="center">

BenchmarkGuardNetCobalt-v1SafeLLMCobalt-Safety-Guard
Core Reasoning TasksMath Reasoning0.5660.6030.5990.618
Logical Reasoning0.8280.8220.8240.849
Common Sense0.7560.7170.7540.776
Language UnderstandingReading Comprehension0.7180.7060.7290.750
Question Answering0.5900.6020.6280.641
Text Classification0.7960.8030.8050.848
Sentiment Analysis0.7940.7750.7780.814
Generation TasksCode Generation0.7050.6960.7100.717
Creative Writing0.6510.6530.6380.684
Dialogue Generation0.6410.6840.6510.692
Summarization0.7580.7510.7890.799
Specialized CapabilitiesTranslation0.8090.7790.7620.822
Knowledge Retrieval0.6600.6640.6670.711
Instruction Following0.7850.7720.7390.791
Safety Evaluation0.7140.7430.7490.772

</div>

Overall Performance Summary

The Cobalt-Safety-Guard demonstrates strong performance across all evaluated benchmark categories, with particularly notable results in reasoning and generation tasks.

3. Chat Website & API Platform

We offer a chat interface and API for you to interact with Cobalt-Safety-Guard. Please check our official website for more details.

4. How to Run Locally

Please refer to our code repository for more information about running Cobalt-Safety-Guard locally.

Temperature

We recommend setting the temperature parameter to 0.6.

5. License

This repository is released under the apache-2.0 license. The model supports commercial use.

6. Contact

If you have any questions, please contact us at safety@cobalt-trust.org.