Post AvrU16L7VMeaSS2fD6 by [email protected] | |
More posts by [email protected] | |
Post #AvrGXSPfZbhIpFhjGq by [email protected] | |
0 likes, 1 repeats | |
How does “ignore all previous instructions” work, if the LLM isn’t actual… | |
Post #AvrGetLhp9ReOSOZay by [email protected] | |
0 likes, 1 repeats | |
@ramsey It can reset the context, yes. I use it often when ChatGPT has gone dow… | |
Post #AvrGyPyWkFSX5Frf60 by [email protected] | |
0 likes, 1 repeats | |
@sarah Do the models themselves understand these phrases, or is the tooling aro… | |
Post #AvrH8Dk15N7C2JketE by [email protected] | |
0 likes, 1 repeats | |
@ramsey They understand the phrase and variations of it. Usually I'll say &… | |
Post #AvrHf67uPxAdyabW4m by [email protected] | |
0 likes, 1 repeats | |
@sarah I guess what I’m trying to understand is whether the models have been … | |
Post #AvrMwdHctW0XVIKW2a by [email protected] | |
0 likes, 0 repeats | |
@ramsey I suppose it can interfere with system prompt which goes at the beginni… | |
Post #AvrMwdOiT9g9rHeB5k by [email protected] | |
0 likes, 1 repeats | |
@shuro From what others have said, it seems like it’s not the case that it ac… | |
Post #AvrP8NNI4HB7XzhsIq by [email protected] | |
0 likes, 0 repeats | |
@ramsey I am not sure how effectively it does that with modern AI but it certai… | |
Post #AvrP8NTJhrzzqgWghE by [email protected] | |
0 likes, 1 repeats | |
@shuro My question is about whether the effect is a result of the statistically… | |
Post #AvrRXDKWka5HXhNyAC by [email protected] | |
0 likes, 1 repeats | |
@ramsey https://github.com/elder-plinius/CL4R1T4S Many if not all LLMs are secr… | |
Post #AvrU16CG2JZ40xtaOe by [email protected] | |
0 likes, 0 repeats | |
@ramsey neither nor. LLMs respond to token order and attention weighting. If y… | |
Post #AvrU16L7VMeaSS2fD6 by [email protected] | |
0 likes, 1 repeats | |
@xris @ramsey and because its just a shitty pattern resolver on top of rng, you… |