Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

DeepSeek-V4-Flash means LLM steering is interesting again

seangoedecke.com

Post reply on HN
Print view View original

21–30 of 84 posts

Re: DeepSeek-V4-Flash means LLM steering is interesting again

#21

Post by andai »

NitpickLawyer wrote: post #18
cyanydeez wrote: post #6

Earlier quoted context omitted.

not sure why youre fixed on censoring. if we invert your POV censoring includes not reporting falsehoods "vaccines are harmful". Science and logic often tackle these subject via censoring, but a model given a equal sampling of Internet, would think vacinnes are harmful. a less naive correction would censor this problematic context. so im cofised as to why you think unmasking whatever bias you think is censored will r…

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…

Anthropic mentioned explicitly making an effort to make Opus 4.7 worse at cybersecurity tasks because the last few generations have been getting too good at them.

So they're trying to improve the model's general intelligence while selectively making it worse in one area.

Re: DeepSeek-V4-Flash means LLM steering is interesting again

#22

Post by andai »

tekne wrote: post #16
cyanydeez wrote: post #6

Earlier quoted context omitted.

not sure why youre fixed on censoring. if we invert your POV censoring includes not reporting falsehoods "vaccines are harmful". Science and logic often tackle these subject via censoring, but a model given a equal sampling of Internet, would think vacinnes are harmful. a less naive correction would censor this problematic context. so im cofised as to why you think unmasking whatever bias you think is censored will r…

So I need to actually check whether these actually end up on separate vectors in current models -- but as a human, there's a huge behavioural difference in: - When doing this task, I should do A and not B - I should refuse to help with this task The former is learning the user's preferences in how to succeed at the task; the latter is determining when to go against the user's chosen task. Your example: - "Are vaccine…

I'm reminded of the emergent misalignment paper, where a model fine-tunes to produce insecure source code would also reliably respond in evil ways to general requests.

e.g. you'd ask it for a cookie recipe and it would add poison to the recipe.

I understood that to be "there was a single neuron "don't be evil" which got inverted" but I'm not sure what it really looks like. (e.g. adding obvious exploits to source code is similar to adding poison to a recipe)

Re: DeepSeek-V4-Flash means LLM steering is interesting again

#23

Post by zozbot234 »

NitpickLawyer wrote: post #18
cyanydeez wrote: post #6

Earlier quoted context omitted.

not sure why youre fixed on censoring. if we invert your POV censoring includes not reporting falsehoods "vaccines are harmful". Science and logic often tackle these subject via censoring, but a model given a equal sampling of Internet, would think vacinnes are harmful. a less naive correction would censor this problematic context. so im cofised as to why you think unmasking whatever bias you think is censored will r…

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.

Re: DeepSeek-V4-Flash means LLM steering is interesting again

#24

Post by bel8 »

Great article but I'm confused on one thing.

The article claims steering only works in local models, but GitHub Copilot has a "steer with message" feature where I can course correct mid execution. I use it often.

I think these are different kinds of steering right? Agent steering probably inserts another user message between the harnesses own ping-pong between harness and the LLM.

- https://docs.github.com/en/copilot/how-tos/copilot-cli/use-c...

- https://docs.github.com/en/copilot/how-tos/copilot-sdk/use-c...

Re: DeepSeek-V4-Flash means LLM steering is interesting again

#26

Post by zozbot234 »

bel8 wrote: post #24

Great article but I'm confused on one thing. The article claims steering only works in local models, but GitHub Copilot has a "steer with message" feature where I can course correct mid execution. I use it often. I think these are different kinds of steering right? Agent steering probably inserts another user message between the harnesses own ping-pong between harness and the LLM. - https://docs.github.com/en/copilot…

Different kind of steering, that's just injecting text into the model's natural language thinking output or something very similar. You can do a middle ground though by using Anthropic's NLA work to look at the natural language rendition of a model's activations at a particular layer, edit the text and convert it back into completely different activations.

Re: DeepSeek-V4-Flash means LLM steering is interesting again

#27

Post by antirez »

amelius wrote: post #25

Sounds more like something for DL research than something you might want to use in practice.

Nope, with the anti-refusal vector loaded you can ask many things for instance related to computer security and if you want to learn, it is a lot better of a model that continuously says you "I can't help you with this problematic request".

Re: DeepSeek-V4-Flash means LLM steering is interesting again

#28

Post by gpugreg »

NitpickLawyer wrote: post #18
cyanydeez wrote: post #6

Earlier quoted context omitted.

not sure why youre fixed on censoring. if we invert your POV censoring includes not reporting falsehoods "vaccines are harmful". Science and logic often tackle these subject via censoring, but a model given a equal sampling of Internet, would think vacinnes are harmful. a less naive correction would censor this problematic context. so im cofised as to why you think unmasking whatever bias you think is censored will r…

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 a vet why my dog looks sick, who then proceeds to cut it open to check. Flagship models usually do not do that without some convincing, but it sure is effective.

We will need much better sandboxes when less restricted models become more common. I can already see them hammering out 0-days when they are prompted to do some task that usually requires root.

Re: DeepSeek-V4-Flash means LLM steering is interesting again

#29

Post by bel8 »

zozbot234 wrote: post #26
bel8 wrote: post #24

Great article but I'm confused on one thing. The article claims steering only works in local models, but GitHub Copilot has a "steer with message" feature where I can course correct mid execution. I use it often. I think these are different kinds of steering right? Agent steering probably inserts another user message between the harnesses own ping-pong between harness and the LLM. - https://docs.github.com/en/copilot…

Different kind of steering, that's just injecting text into the model's natural language thinking output or something very similar. You can do a middle ground though by using Anthropic's NLA work to look at the natural language rendition of a model's activations at a particular layer, edit the text and convert it back into completely different activations.

Ahh I see. Thanks for the clarification.

Re: DeepSeek-V4-Flash means LLM steering is interesting again

#30

Post by surgical_fire »

cyanydeez wrote: post #6
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…

not sure why youre fixed on censoring. if we invert your POV censoring includes not reporting falsehoods "vaccines are harmful". Science and logic often tackle these subject via censoring, but a model given a equal sampling of Internet, would think vacinnes are harmful. a less naive correction would censor this problematic context. so im cofised as to why you think unmasking whatever bias you think is censored will r…

This is something difficult to handle properly.

I think it is useful to turn off censoring if you need.

When I am researching something, I likely want proper information. If I am looking up information on vaccines, I don't want information that crackpots spread online on chips on vaccines and how 5g will kill the vaccinated, or how it is somehow connected with Bill Gates spreading meat allergies through drones raining ticks on unsuspecting people.

On the other hand, if I am actively looking up crazy bullshit information (perhaps I want some entertainment), I should be able to read it.

Post reply on HN