A recent paper from Apple about reasoning deficits has been widely reposted as “LLMs Can’t Reason”. The study claims to demonstrate significant limitations in the reasoning capabilities of large language models (LLMs). Gary Marcus, author of a “Forbes 7 Must Read Books in AI”, railed: “There is just no way can you build reliable agents on this f... Read more 20 Oct 2024 - 3 minute read
As I was compiling bug reports for a Vision Language Model vendor, I found the need to redact images without JPEG recompression: simply re-saving a particular sample image that had originally triggered a repitition loop with the MLLM changed the image in such a way that I could not reproduce the problem. Solution: asenior/Jpeg-Redaction-Library.... Read more 15 Oct 2024 - 1 minute read
The community is porting the “Shrek Sampler” to different hardware (MLX) and Transformer architectures, just days after Entropix was first released. Nice visualization from the MLX port: 9.9 vs. 9.11. Pierre-Carl Langlais, Co-founder of Pleias.fr, posted his Colab notebook that runs entropix with Smollm-360M (the original release uses Llama 3.2 ... Read more 13 Oct 2024 - less than 1 minute read
The State of AI report 2024 is available. Some notes: their definition of “Agent” seems reasonable: “AI Agent: an AI-powered system that can take actions in an environment. For example, an LLM that has access to a suite of tools and has to decide which one to use in order to accomplish a task that it has been prompted to do.”. Jaana Dogan, Pr... Read more 10 Oct 2024 - 2 minute read
Particularly when working on a source code project, users find the need to submit several files to an LLM at once. Several software project to aid with this have sprung into existence. A thread on HackerNews listed at least seven different ones, one of them being Repopack. I like to avoid installing non-essential software, so found it useful to... Read more 06 Oct 2024 - 1 minute read