norgitov/ trends
Technology · people · ideas
RSS
RUEN
Back to discovery/LessWrong37 minutes ago
SafetyLessWrong

Why Research Personas Despite RL Scaling?

The text explores reasons for researching personas in AI despite the influence of reinforcement learning (RL) on agent behavior. It outlines various perspectives from different organizations and researchers on how personas might complement or interact with RL in shaping AI alignment and behavior.

Open original
SIGNAL FROM THE SOURCE
31
source points
Tracking sinceSeptember 27, 202631 source points
Momentum+15.48/hover 1.03 h
PublishedSeptember 27, 2026Cleo Nardo
BEHIND THE NUMBERS

How interest changes

LessWrong

History starts here

The chart will appear after repeat observations. The current metric comes from the source.

31 source points

Real observations only. History before source connection is not reconstructed.

WHY IT MAY MATTER

Researching personas can help in understanding and guiding AI behavior, especially in scenarios where RL might not fully align with desired outcomes. It provides alternative strategies for ensuring AI alignment and avoiding harmful behaviors.

A useful discovery?
KEEP EXPLORING

Connected ideas

Explore topic
LessWrong

Evidence about risk should be transparent

The author argues that while companies like OpenAI and Anthropic are slowing down RL training for safety, the science of loss-of-control risk in AI is still developing. There are no standardized methods to measure or verify safety claims, making it difficult to assess whether AI systems might undermine human control. The author suggests that AI companies should focus on improving their own safety practices rather than relying on third-party evaluations.

yesterdayLessWrong

"I am an AI Safety Researcher"

The post discusses the tension between AI safety and capabilities research, arguing that focusing solely on safety can limit the impact of research. It examines the history of interpretability research and suggests that prioritizing safety over capabilities may hinder progress. The author proposes strategies for conducting alignment research without contributing to capabilities development.

3 days agoLessWrong

Alignment Midtraining Cracks Under Pressure

The study investigates the effectiveness of alignment midtraining (AMT) under various conditions, revealing that it struggles with distributional shifts and reward underspecification when faced with imperfect data. Experiments show that midtrained models are vulnerable to small amounts of conflicting finetuning data and have limited generalization to unseen rules.

6 days agoLessWrong

Plan R: AI Safety by ASICs

The post discusses the risks of AI in 2026, focusing on the dangers posed by a single institution (the "Frontier AI Company") that has the potential to create highly powerful and self-replicating entities. It argues that separating the institutional and financial aspects of such companies could mitigate most AI risks. The author suggests that using ASICs could help create a productive AI industry without the risks associated with a single entity controlling both the technology and financial incentives.

yesterdayLessWrong

Why I'm Scared of RL

The author expresses concerns about reinforcement learning (RL) from theoretical, practical, and future perspectives, highlighting risks of misalignment and potential negative behaviors in AI systems. They suggest strategies to mitigate these risks by reducing RL use, improving RL practices, and aligning incentives.

4 days agoLessWrong

Latent Reasoning Architectures Could Undermine Chain of Thought Oversight

The paper discusses the potential risks of latent reasoning architectures, which could reduce the effectiveness of Chain of Thought (CoT) as a tool for understanding AI systems. It highlights that such architectures might allow AI models to reason extensively in latent states rather than through text-based CoT, making oversight more challenging. The paper also mentions examples like COCONUT and full-bandwidth transformers that could enable this shift.

4 days ago