This is a link post.

Produced as part of the MATS Summer 2024 program, under the mentorship of Alex Turner (TurnTrout).

A few weeks ago, I stumbled across a very weird fact: it is possible to find multiple steering vectors in a language model that activate very similar behaviors while all being orthogonal. This was pretty surprising to me and to some people that I talked to, so I decided to write a post about it. I don't currently have the bandwidth to investigate this much more, so I'm just putting this post and the code up.

I'll first discuss how I found these orthogonal steering vectors, then share some results. Finally, I'll discuss some possible explanations for what is happening.

Methodology

My work here builds upon Mechanistically Eliciting Latent Behaviors in Language Models (MELBO). I use MELBO to find steering vectors. Once I have a MELBO vector, I then [...]

---

Outline:

(00:50) Methodology

(03:30) Results

(08:46) Hypotheses for what is happening

(11:33) Conclusion

(12:11) Appendix: other orthogonal vectors

(12:22) becomes an alien species vector results

(12:43) STEM problem vector results

(13:07) jailbreak vector results

The original text contained 9 images which were described by AI.

---

“I found >800 orthogonal ‘write code’ steering vectors” by Jacob G-W, TurnTrout

LessWrong (30+ Karma)