Abel Jansma
A diagonal attack for LLM truth probes shows why no probe on a language model’s embedding space can pin down truth. A linear dream Modern LLMs famously encode input texts as vectors in some embedding space. One of the most satisfying discoveries about LLMs is that many natural concepts,...
Abel Jansma
In a previous post, I proposed to study complex systems through a mereological lens by applying the Möbius inversion theorem. This has become my favourite theorem by now, because I think it allows you to do integration and differentiation on observables in complex systems where this was not...
Abel Jansma
UPDATE: I have now found an explicit formula for the Möbius function on the redundancy lattice. This will appear in a preprint soon. I calculated the full Möbius function on the lattice of powerset antichains of 2, 3, and 4 variables. As far as I can tell, these results have not been shared...
Abel Jansma
In my PhD thesis, I developed a method and software package to infer interactions and cell states from gene expression data. Both the software and the research are now published! The software is called Stator, and comprises the Stator Nextflow pipeline, as well as a bespoke Shiny app developed by...
Abel Jansma
Generative models, in particular energy-based models, are often used to sample from conditional distributions–a process known as inpainting. One of the most fundamental kinds of generative energy-based models is called a restricted Boltzmann machine (RBM), which is essentially a bipartite, glassy...
1ucasvb's lab
Can you flatten a sphere? The answer is NO, you can not. This is why all map projections are innacurate and distorted, requiring some form of compromise between how accurate the angles, distances and areas in a globe are represented. This is all due to Gauss’s Theorema Egregium, which dictates...