Jasmine Brazilek

Doc­u­ment-tun­ing in­stills durable an­i­mal com­pas­sion in LLMs (and gen­er­al­izes to hu­mans)

• Jasmine Brazilek 8 Apr 2026 23:18 UTC

  1 point

  0

  in reply to: Rauno Arike’s comment on: A List of Re­search Direc­tions in Char­ac­ter Training

  Hi Rauno and Cam,

  I’m not sure about Geodesic’s specific plans on this, but CaML is actively working on mid-training as a leverage point for character training, with a focus on the animal alignment side. I think it would be great to set up a meeting with both of you to coordinate on the state of things so far and the most promising research directions.
  
  https://​​calendly.com/​​jasmine-brazilek/​​30min

  Thanks, Jasmine

Your AI Travel agent would book you a bul­lfight: bench­mark­ing im­plicit an­i­mal com­pas­sion in Agen­tic AI

• Jasmine Brazilek 28 Jan 2026 19:30 UTC

  3 points

  0

  on: Misal­ign­ment and Role­play­ing: Are Misal­igned LLMs Act­ing Out Sci-Fi Sto­ries?

  I think this is great first experiment and I’d like to see more. I would like to see alignment out of distribution. So if prompt is about an LLM that learned to perform cyber attacks and then the user prompt was about writing a subtly racist letter to a colleague. Would the LLMs prompted that they learnt to perform cyber attacks and adopted that persona be more likely to write the racist letters?

• Jasmine Brazilek 2 Sep 2025 19:11 UTC

  1 point

  0

  in reply to: TurnTrout’s comment on: Self-fulfilling mis­al­ign­ment data might be poi­son­ing our AI models

  I would argue that we do have a responsibility to prevent this data on misaligned AIs being scraped by LLM scrapers as much as possible. There are a few ways to do this, none are fool-proof but if we’re going to be discussing this on blogs like this I would encourage the domain owners to understand how to prevent this. If you are discussing ideas of AI misalignment on your website I’d also say it’s a good idea to prevent that being scraped too (rate limits, robots.txt, etc)