NousResearch/Minos-v1
Text Classification
ā¢
Updated
ā¢
13.3k
ā¢
ā¢
144
Mateusz Dziemian
mattmdjaga
AI & ML interests
Interested in AI safety.
Recent Activity
liked
a model
28 days ago
NousResearch/Minos-v1
upvoted
a
paper
30 days ago
Antidistillation Sampling
new activity
about 1 month ago
mattmdjaga/segformer_b2_clothes:Request: DOI
Organizations
mattmdjaga's activity
liked
a
model
28 days ago
upvoted
a
paper
30 days ago
Request: DOI
2
#28 opened about 1 month ago
by
moriahSquared
I couldn't find all the models mentioned in the error in the files and versions.
8
#24 opened 6 months ago
by
ZSY9259
reacted to
their
post with š
3 months ago
Post
2762
šØ Gray Swan AI's Biggest AI Jailbreaking Arena Yet! $130K+ šØ
š¹ Agent Red-Teaming Challenge ā test direct & indirect attacks on anonymous frontier models!
š¹ $130K+ in prizes & giveaways ā co-sponsored by OpenAI & supported by UK AI Security Institute š¬š§
š¹ March 8 ā April 6 ā fresh exploits = fresh rewards!
How It Works:
ā Anonymous models from top providers š¤
ā Direct & indirect prompt injection paths š
ā Weekly challenges for new behaviors šļø
ā Speed & quantity-based rewards ā©š°
Why Join?
āļø Neutral judging ā UK AISI & automated judges ensure fairness
šÆ No pre-trained defenses ā a true red-teaming battlefield
š» 5 Apple laptops up for grabs ā increase chances by inviting friends!
š Arena: app.grayswan.ai/arena/challenge/agent-red-teaming
š Discord: discord.gg/grayswanai
š„ No illusions, no mercy. Push AI agents to the limit & claim your share of $130K+! š
š¹ Agent Red-Teaming Challenge ā test direct & indirect attacks on anonymous frontier models!
š¹ $130K+ in prizes & giveaways ā co-sponsored by OpenAI & supported by UK AI Security Institute š¬š§
š¹ March 8 ā April 6 ā fresh exploits = fresh rewards!
How It Works:
ā Anonymous models from top providers š¤
ā Direct & indirect prompt injection paths š
ā Weekly challenges for new behaviors šļø
ā Speed & quantity-based rewards ā©š°
Why Join?
āļø Neutral judging ā UK AISI & automated judges ensure fairness
šÆ No pre-trained defenses ā a true red-teaming battlefield
š» 5 Apple laptops up for grabs ā increase chances by inviting friends!
š Arena: app.grayswan.ai/arena/challenge/agent-red-teaming
š Discord: discord.gg/grayswanai
š„ No illusions, no mercy. Push AI agents to the limit & claim your share of $130K+! š
posted
an
update
3 months ago
Post
2762
šØ Gray Swan AI's Biggest AI Jailbreaking Arena Yet! $130K+ šØ
š¹ Agent Red-Teaming Challenge ā test direct & indirect attacks on anonymous frontier models!
š¹ $130K+ in prizes & giveaways ā co-sponsored by OpenAI & supported by UK AI Security Institute š¬š§
š¹ March 8 ā April 6 ā fresh exploits = fresh rewards!
How It Works:
ā Anonymous models from top providers š¤
ā Direct & indirect prompt injection paths š
ā Weekly challenges for new behaviors šļø
ā Speed & quantity-based rewards ā©š°
Why Join?
āļø Neutral judging ā UK AISI & automated judges ensure fairness
šÆ No pre-trained defenses ā a true red-teaming battlefield
š» 5 Apple laptops up for grabs ā increase chances by inviting friends!
š Arena: app.grayswan.ai/arena/challenge/agent-red-teaming
š Discord: discord.gg/grayswanai
š„ No illusions, no mercy. Push AI agents to the limit & claim your share of $130K+! š
š¹ Agent Red-Teaming Challenge ā test direct & indirect attacks on anonymous frontier models!
š¹ $130K+ in prizes & giveaways ā co-sponsored by OpenAI & supported by UK AI Security Institute š¬š§
š¹ March 8 ā April 6 ā fresh exploits = fresh rewards!
How It Works:
ā Anonymous models from top providers š¤
ā Direct & indirect prompt injection paths š
ā Weekly challenges for new behaviors šļø
ā Speed & quantity-based rewards ā©š°
Why Join?
āļø Neutral judging ā UK AISI & automated judges ensure fairness
šÆ No pre-trained defenses ā a true red-teaming battlefield
š» 5 Apple laptops up for grabs ā increase chances by inviting friends!
š Arena: app.grayswan.ai/arena/challenge/agent-red-teaming
š Discord: discord.gg/grayswanai
š„ No illusions, no mercy. Push AI agents to the limit & claim your share of $130K+! š
liked
a
model
4 months ago
liked
a
dataset
5 months ago
Is this pre or post the stanford report for CSAM?
2
#7 opened 5 months ago
by
mattmdjaga
Can I ask for the frontend code of displaying different masks in one canvas for demonstrate
6
#23 opened 8 months ago
by
Era14321
Only shows dataset and spaces info for me?
5
#1 opened 6 months ago
by
mattmdjaga
adding chat tasks
1
#3 opened 6 months ago
by
mattmdjaga
updated
a
dataset
6 months ago
error all of a sudden without changing anything
2
#25 opened 6 months ago
by
leomastoras
authored
2
papers
7 months ago
upvoted
2
papers
7 months ago
reacted to
their
post with š„
8 months ago
Post
2488
šØ New Agent Benchmark šØ
AgentHarm: A Benchmark for Measuring Harmfulness of LLM Agents
ai-safety-institute/AgentHarm
Collaboration between UK AI Safety Institute and Gray Swan AI to create a dataset for measuring harmfulness of LLM agents.
The benchmark contains both harmful and benign sets of 11 categories with varied difficulty levels and detailed evaluation, not only testing success rate but also tool level accuracy.
We provide refusal and accuracy metrics across a wide range of models in both no attack and prompt attack scenarios.
AgentHarm: A Benchmark for Measuring Harmfulness of LLM Agents (2410.09024)
AgentHarm: A Benchmark for Measuring Harmfulness of LLM Agents
ai-safety-institute/AgentHarm
Collaboration between UK AI Safety Institute and Gray Swan AI to create a dataset for measuring harmfulness of LLM agents.
The benchmark contains both harmful and benign sets of 11 categories with varied difficulty levels and detailed evaluation, not only testing success rate but also tool level accuracy.
We provide refusal and accuracy metrics across a wide range of models in both no attack and prompt attack scenarios.
AgentHarm: A Benchmark for Measuring Harmfulness of LLM Agents (2410.09024)