Proposal: Safeguarding Against Jailbreaking Through Iterative Multi-TurnTesting

jacquesallen31 Jan 2025 23:00 UTC

3 points

0 comments8 min readLW link

The Failed Strategy of Artificial Intelligence Doomers

Ben Pace31 Jan 2025 18:56 UTC

76 points

21 comments4 min readLW link

(www.palladiummag.com)

Safe Search is off: root causes of AI catastrophic risks

Jemal Young31 Jan 2025 18:22 UTC

2 points

0 comments3 min readLW link

5,000 calories of peanut butter every week for 3 years straight

Declan Molony31 Jan 2025 17:29 UTC

8 points

4 comments1 min readLW link

Will alignment-faking Claude accept a deal to reveal its misalignment?

ryan_greenblatt and Kyle Fish

31 Jan 2025 16:49 UTC

127 points

7 comments12 min readLW link

Some articles in “International Security” that I enjoyed

Buck31 Jan 2025 16:23 UTC

32 points

1 comment4 min readLW link

[Question] How do biological or spiking neural networks learn?

Dom Polsinelli31 Jan 2025 16:03 UTC

1 point

0 comments2 min readLW link

Defense Against the Dark Prompts: Mitigating Best-of-N Jailbreaking with Prompt Evaluation

Stuart_Armstrong and rgorman

31 Jan 2025 15:36 UTC

18 points

1 comment2 min readLW link

Catastrophe through Chaos

Marius Hobbhahn31 Jan 2025 14:19 UTC

80 points

6 comments12 min readLW link

Review: The Lathe of Heaven

dr_s31 Jan 2025 8:10 UTC

21 points

0 comments8 min readLW link

[Question] Is weak-to-strong generalization an alignment technique?

cloud31 Jan 2025 7:13 UTC

12 points

0 comments2 min readLW link

Takeaways from sketching a control safety case

joshc31 Jan 2025 4:43 UTC

28 points

0 comments3 min readLW link

(redwoodresearch.substack.com)

Steering Gemini with BiDPO

TurnTrout31 Jan 2025 2:37 UTC

79 points

3 comments1 min readLW link

(turntrout.com)

In response to critiques of Guaranteed Safe AI

Nora_Ammann31 Jan 2025 1:43 UTC

33 points

2 comments26 min readLW link

Proposal for a Form of Conditional Supplemental Income (CSI) in a Post-Work World

sweenesm31 Jan 2025 1:00 UTC

2 points

0 comments3 min readLW link

Outlaw Code

scarcegreengrass30 Jan 2025 23:41 UTC

6 points

0 comments2 min readLW link

Can someone, anyone, make superintelligence a more concrete concept?

Ori Nagel30 Jan 2025 23:25 UTC

2 points

4 comments4 min readLW link

What’s Behind the SynBio Bust?

sarahconstantin30 Jan 2025 22:30 UTC

46 points

5 comments6 min readLW link

(sarahconstantin.substack.com)

The future of humanity is in management

jasoncrawford30 Jan 2025 22:14 UTC

1 point

2 comments13 min readLW link

(newsletter.rootsofprogress.org)

A sketch of an AI control safety case

Tomek Korbak, joshc, Benjamin Hilton, Buck and Geoffrey Irving

30 Jan 2025 17:28 UTC

50 points

0 comments5 min readLW link