My first PhD paper!🎉We learn *diffusion* models for code generation that learn to directly *edit* syntax trees of programs. The result is a system that can incrementally write code, see the execution output, and debug it. 🧵1/n
π-0.5 is here, and it can generalize to new homes! Some fun experiments with my colleagues at @physical_int, introducing π-0.5 (“pi oh five”). Our new VLA can put dishes in the sink, clean up spills and do all this in homes that it was not trained in🧵👇
@JeremyNguyenPhD I'm trying to figure out a good way to share this, since running GPUs is pretty expensive. Though my plans were more so to turn this from a demo to a more polished toy/game 😊
@xf1280 I experimented with a bunch of image to 3D models. In this video I'm using fast3d mainly because of very low latency, though in my experiments other models like hunyuan3d and trellis gave better quality meshes.
@alexanderchen Thanks Alex! The "system" prompt I wrote specifies that the model should largely follow the user sketch, but if it's a particularly bad sketch, the model is allowed to be creative. It would be so cool to include that more as a slider for user control ✨
LMs can generalize to implications of facts they are finetuned on. But what mechanisms enable this, and how are these mechanisms learned in pretraining? We develop conceptual and empirical tools for studying these qns. 🧵
Can interpretability help defend LLMs? We find we can reshape activations while preserving a model’s behavior. This lets us attack latent-space defenses, from SAEs and probes to Circuit Breakers. We can attack so precisely that we make a harmfulness probe output this QR code. 🧵
I'll be at NeurIPS, let me know if you want to catch up or chat about program synthesis, world models, neurosymbolic, search, probabilistic programming, or mourning the loss of King Da Ka.
I am currently holding my dad's cryopreserved brain tumor samples in hopes of creating a personalized vaccine for immunotherapy. However, there are some critical and time-sensitive questions in the attached post: https://t.co/1haayeNsa0
This is time-sensitive so would appreciate any DMs/RTs.
@EmilevanKrieken I think it has a lot of synergies with GFlowNets (which we mention in the paper) and one of our baseline methods (REPL Flow) is a mix between Ellis et. al. reimagined as a GFlowNet.
My first PhD paper!🎉We learn *diffusion* models for code generation that learn to directly *edit* syntax trees of programs. The result is a system that can incrementally write code, see the execution output, and debug it. 🧵1/n
@EmilevanKrieken In our current mutation scheme, the expression can get longer or shorter at roughly the same probability, so not sure about the limiting distribution. Anecdotally we noticed that if we noise the program some number of times, the programs resemble just random programs.