Cameron R. Wolfe, Ph.D.
Director of AI @RebuyEngine • Founder @ Deep (Learning) Focus • PhD @optimalab1 • I make AI understandable
View on 𝕏Threads
The volume of LLM research being released is staggering. Although there are too many new papers for any one person to read, this work can be largely distilled into a much smaller s...
LLaMA-2 outlines the remaining limitations of open-source language models well. Put simply, the gap in performance between open-source and proprietary LLMs is largely due to the qu...
When querying a language model, there are many ways to tweak the randomness of the model’s generations. One popular method is via the “top_k” setting, which works in tandem with th...
There are three primary ways in which language models learn. Let’s quickly go over them and how they are different from each other… 🧵 [1/7] https://t.co/iaInEBqAIe
Foundation models are a popular topic in AI research. However, task-specific fine-tuning outperforms zero/few-shot learning with foundation models in most cases. Specialized models...
Prompt engineering for language models usually involves tweaking the wording or structure of a prompt. But, recent research has explored automated prompt engineering via continuous...
I’m currently writing a survey/overview of important prompt engineering tactics for my newsletter. Here’s my top-5 findings so far… 🧵 [1/7] https://t.co/I6m2mpepQs
Large Language Models (LLMs) are notoriously bad at solving reasoning-based tasks. However, we can drastically improve their reasoning performance using simple techniques that requ...
Nearly all recently-proposed large language models (LLMs) are based upon the decoder-only transformer architecture. But, is this always the best architecture to use? It depends… 🧵...
Following the release of LLaMA, we saw a rapid explosion of open-source research on large language models (LLMs). Here are the three most notable model releases during this time… 🧵...
As Large Language Models (LLMs) improve in quality, evaluating them becomes more difficult. Recent models are so good that even humans struggle to discern differences in quality. L...
Each “block” of a large language model (LLM) is comprised of self-attention and a feed-forward transformation. However, the exact self-attention variant used by LLMs is masked, mul...