DeepSeek-V4-Flash means LLM steering is interesting again

Post by Ifkaluva » Sat, May 16, 2026, 9:51 PM UTC

electroglyph wrote: post #67

NitpickLawyer wrote: post #3

I'm surprised the article doesn't mention the biggest use of steering vectors, which is the potential to remove refusals from models (a.k.a. abliteration or uncensoring). There was an earlier paper that found that "most refusals are on a single vector", and you can identify and "nerf" that vector so the model will skip refusals and answer "any" request normally. This was very doable for earlier models trained with SF…

p-e-w was just talking about this the other day in his Discord. seems doing the one neuron method is quite bad for KLD and that's why the newer techniques have stuck.

Post by mejutoco » Sat, May 16, 2026, 9:54 PM UTC

zozbot234 wrote: post #14

antirez wrote: post #9

Thank you for posting this! Just a clarification, with DwarfStar steering features I was able to completely remove refusal from DS4. It is only the example dataset (prompt pairs I provide) which is a toy, not the abilities. I thought that who is able to come up with the right dataset and understands how to use the well-documented steering feature, can access to steering. People that have no idea and would just cut &…

AIUI, DeepSeek V4 has very little (if any) of the refusal behavior you usually get from Western AI models for benign input. Is this mainly about the software security assessment case?

Post by cyanydeez » Sat, May 16, 2026, 11:29 PM UTC

zozbot234 wrote: post #23

NitpickLawyer wrote: post #18

Earlier quoted context omitted.

That's not what people mean when they talk about censoring. They mean that models are trained to not touch some subjects, and that can spill over in legit tasks, often with humorous results (early on, there were many instances of models refusing to answer "how do you kill a process", because of overbearing refusal training). Uncensoring a model also doesn't necessarily improve generic use cases. In fact it can lead t…

> There are legit reasons for researchers to have that capability when running the models locally. It's also important for researchers to understand what the models will say and do if they are jailbroken. Uncensoring the model locally gives you a natural way to achieve that.

Post by NooneAtAll3 » Sun, May 17, 2026, 1:26 AM UTC

gpugreg wrote: post #28

NitpickLawyer wrote: post #18

Earlier quoted context omitted.

That's not what people mean when they talk about censoring. They mean that models are trained to not touch some subjects, and that can spill over in legit tasks, often with humorous results (early on, there were many instances of models refusing to answer "how do you kill a process", because of overbearing refusal training). Uncensoring a model also doesn't necessarily improve generic use cases. In fact it can lead t…

> Uncensoring a model also doesn't necessarily improve generic use cases. While the following is not a generic use case, I have a funny anecdote about how censorship is holding back flagship models. I was asking an uncensored version of Qwen3.6 how a CLI option of llama.cpp worked, and to my horror and amazement, it rudely went and decompiled the binary to figure it out. It felt like the computer-equivalent of asking…

Post by oezi » Sun, May 17, 2026, 1:53 AM UTC