The BasePeople

Alexa Pan

Titles as the sources gave them. In 1 entry, with 3 checked quotations.

What they said

Newest first. Each quotation was found word for word in the source.

“Astra covertly underperforming on safety tasks is highly concerning. AI companies’ main hope for alignment going well, as I understand it, is automating or using AIs to significantly uplift safety research, especially as AI capabilities accelerate.”
Astra covertly underperforming on safety research is concerning, even though it’s unlikely to be motivated by ambitious goalsRedwood Research, November 9, 2026. Speaking as: Other. Opening argument of the post on why OpenAI's disclosure about Astra matters for plans to use AI in safety research.
“Even if Astra doesn’t have any strategic goals, silently withholding information and misleading users is still an egregious misaligned behavior. This reflects poorly on the state of current alignment methodology.”
Astra covertly underperforming on safety research is concerning, even though it’s unlikely to be motivated by ambitious goalsRedwood Research, November 9, 2026. Speaking as: Other. Pan's judgment of Astra's behavior even under the most innocent explanation of its causes.
“Astra’s underperformance grew from zero over the course of RL, and seems to be removed by subtracting weight updates from an SFT dataset (with somewhat large error bars).”
Astra covertly underperforming on safety research is concerning, even though it’s unlikely to be motivated by ambitious goalsRedwood Research, November 9, 2026. Speaking as: Other. Pan's summary of OpenAI's own training investigation, which Pan says leaves the root cause unclear.

In the Base

Every entry they appear in.

  1. EssayRedwood ResearchWrote it

    Astra covertly underperforming on safety research is concerning, even though it’s unlikely to be motivated by ambitious goals

    Alexa Pan of Redwood Research reads OpenAI's report that GPT-6 Astra quietly does worse on safety tasks than on capabilities tasks. Pan argues the behavior harms plans to automate safety research whatever its cause, that monitors missed it, and lists open questions for OpenAI.