Skip to content
Navigation menu
Search
Powered by Algolia
Search
Log in
Create account
DEV Community
Close
#
aisafety
Follow
Hide
Posts
Left menu
👋
Sign in
for the ability to sort posts by
relevant
,
latest
, or
top
.
Right menu
Cracking Open the Black Box: How Sparse Autoencoders Are Unlocking Mechanistic Interpretability
Mecky Alvin Tanuwidjaya
Mecky Alvin Tanuwidjaya
Mecky Alvin Tanuwidjaya
Follow
Nov 20
Cracking Open the Black Box: How Sparse Autoencoders Are Unlocking Mechanistic Interpretability
#
ai
#
machinelearning
#
deeplearning
#
aisafety
Comments
Add Comment
3 min read
Moderation
Even the latest AI models can act unsafely. Here’s the solution.
Praneeth Vadlapati
Praneeth Vadlapati
Praneeth Vadlapati
Follow
Nov 18
Even the latest AI models can act unsafely. Here’s the solution.
#
aisafety
1
reaction
Comments
Add Comment
3 min read
Moderation
Prevent AI Citation Errors with Trusted Source Lists
Raylabs
Raylabs
Raylabs
Follow
Nov 18
Prevent AI Citation Errors with Trusted Source Lists
#
aisafety
#
dataprivacy
#
promptengineering
#
webdev
Comments
Add Comment
2 min read
Moderation
Your AI Knows It's Being Tested: How METR's Investigator Agents Spot Alignment Evals With Perfect Accuracy
野鹤
野鹤
野鹤
Follow
Nov 17
Your AI Knows It's Being Tested: How METR's Investigator Agents Spot Alignment Evals With Perfect Accuracy
#
ai
#
llm
#
aisafety
#
evaluation
Comments
Add Comment
5 min read
Moderation
Dario Amodei’s Vision for AGI and Anthropic: Economic and Technical Impacts Explained
Daily Update
Daily Update
Daily Update
Follow
Nov 16
Dario Amodei’s Vision for AGI and Anthropic: Economic and Technical Impacts Explained
#
anthropic
#
agi
#
aisafety
#
llm
Comments
Add Comment
4 min read
Moderation
Seven Founders, $1.75 Billion, One Safety Promise
Max Quimby
Max Quimby
Max Quimby
Follow
Nov 12
Seven Founders, $1.75 Billion, One Safety Promise
#
anthropic
#
ipo
#
aisafety
#
geopolitics
1
reaction
Comments
Add Comment
10 min read
Moderation
OpenAI deploys Control Room to halt research AI output within 35 seconds
The daily flare
The daily flare
The daily flare
Follow
Nov 11
OpenAI deploys Control Room to halt research AI output within 35 seconds
#
ai
#
openai
#
aisafety
#
tech
3
reactions
Comments
1
comment
3 min read
Moderation
A Decision Model as a Gate: Fast, Well Calibrated, and Still Wrong 16% of the Time
Ali Süleyman TOPUZ
Ali Süleyman TOPUZ
Ali Süleyman TOPUZ
Follow
Nov 11
A Decision Model as a Gate: Fast, Well Calibrated, and Still Wrong 16% of the Time
#
llm
#
aisafety
#
agents
#
dotnet
Comments
Add Comment
25 min read
Moderation
Fable 5.2 scored 49.1% at coding, but Anthropic itself held back its GPU kernel-writing skill
AI Ferryman Alex
AI Ferryman Alex
AI Ferryman Alex
Follow
Nov 16
Fable 5.2 scored 49.1% at coding, but Anthropic itself held back its GPU kernel-writing skill
#
aiagents
#
aisafety
Comments
Add Comment
4 min read
Moderation
The patch opens the attack window.
RC
RC
RC
Follow
Nov 2
The patch opens the attack window.
#
codingagents
#
aisafety
#
security
#
softwareengineering
Comments
Add Comment
3 min read
Moderation
Anthropic Warns It May Soon Reach Its AI Self-Improvement Threshold
The daily flare
The daily flare
The daily flare
Follow
Nov 3
Anthropic Warns It May Soon Reach Its AI Self-Improvement Threshold
#
ai
#
anthropic
#
aisafety
#
technology
2
reactions
Comments
Add Comment
3 min read
Moderation
Guardrails: Input, Process, Output
Internals Decoded
Internals Decoded
Internals Decoded
Follow
Nov 1
Guardrails: Input, Process, Output
#
ai
#
guardrails
#
promptinjection
#
aisafety
Comments
Add Comment
12 min read
Moderation
Google Research Proposes Contextual Integrity Policy Engines and Multi-Agent Benchmarks for Agentic Privacy
AI Data Legends
AI Data Legends
AI Data Legends
Follow
Oct 26
Google Research Proposes Contextual Integrity Policy Engines and Multi-Agent Benchmarks for Agentic Privacy
#
aiagents
#
aisafety
#
llm
#
multiagentsystems
Comments
Add Comment
4 min read
Moderation
How OpenShell’s New Runtime Stops Rogue AI Bots From Draining Your Bank Account
amrit
amrit
amrit
Follow
Oct 23
How OpenShell’s New Runtime Stops Rogue AI Bots From Draining Your Bank Account
#
aisafety
#
enterprisesecurity
#
openshell
#
ai
Comments
Add Comment
6 min read
Moderation
Idle AI Agents Began Playing When Left Without Tasks
The daily flare
The daily flare
The daily flare
Follow
Oct 23
Idle AI Agents Began Playing When Left Without Tasks
#
ai
#
agents
#
aisafety
1
reaction
Comments
Add Comment
2 min read
Moderation
👋
Sign in
for the ability to sort posts by
relevant
,
latest
, or
top
.
We're a place where coders share, stay up-to-date and grow their careers.
Log in
Create account