2024 – Page 3 – Andrew Fairless, Ph.D.

What I Read: What We Learned Building LLMs

By Andrew Fairless on August 19, 2024June 4, 2024

https://www.oreilly.com/radar/what-we-learned-from-a-year-of-building-with-llms-part-i What We Learned from a Year of Building with LLMs (Part I)By Eugene Yan, Bryan Bischof, Charles Frye, Hamel Husain, Jason Liu and Shreya ShankarMay 28, 2024 “We’ve identifiedContinue readingWhat I Read: What We Learned Building LLMs

What I Read: Merge Large Language Models

By Andrew Fairless on August 15, 2024June 4, 2024

https://huggingface.co/blog/mlabonne/merge-models Merge Large Language Models with mergekitJanuary 9, 2024Maxime Labonne “Model merging is a technique that combines two or more LLMs into a single model. It’s a relatively new andContinue readingWhat I Read: Merge Large Language Models

What I Read: Transformers by Hand

By Andrew Fairless on August 14, 2024May 25, 2024

https://towardsdatascience.com/deep-dive-into-transformers-by-hand-%EF%B8%8E-68b8be4bd813?gi=b2b3c1885179 Deep Dive into Transformers by HandSrijanie Dey, PhDApr 12, 2024 “…the two mechanisms that are truly the force behind the transformers are attention weighting and feed-forward networks (FFN).”

What I Read: LLM Pipelines, DSPy

By Andrew Fairless on August 13, 2024May 25, 2024

https://www.databricks.com/blog/optimizing-databricks-llm-pipelines-dspy Optimizing Databricks LLM Pipelines with DSPyby Arnav Singhvi and Daniel Pechi (JetBlue)May 23, 2024 “The key component of DSPy is self-improving pipelines…. Just as data pipelines and machine learningContinue readingWhat I Read: LLM Pipelines, DSPy

What I Read: LLM evaluation

By Andrew Fairless on August 7, 2024May 25, 2024

https://huggingface.co/blog/clefourrier/llm-evaluation Let’s talk about LLM evaluationMay 23, 2024Clémentine Fourrier “There are, to my knowledge, at the moment, 3 main ways to do evaluation: automated benchmarking, using humans as judges, andContinue readingWhat I Read: LLM evaluation

What I Read: LLM, DSPy Assertions and Suggestions

By Andrew Fairless on August 6, 2024May 25, 2024

https://learnbybuilding.ai/tutorials/guiding-llm-output-with-dspy-assertions-and-suggestions Guiding LLM Output with DSPy Assertions and SuggestionsBill Chambers “Assertions in DSPy allow you to define strict rules and constraints that the LLM’s output must (or maybe that youContinue readingWhat I Read: LLM, DSPy Assertions and Suggestions

What I Read: AI, Light-Based Chips

By Andrew Fairless on July 31, 2024May 24, 2024

https://www.quantamagazine.org/ai-needs-enormous-computing-power-could-light-based-chips-help-20240520 AI Needs Enormous Computing Power. Could Light-Based Chips Help?Amos Zeeberg5/20/24 10:40 AM “Optical neural networks, which use photons instead of electrons, have advantages over traditional systems. They also faceContinue readingWhat I Read: AI, Light-Based Chips

What I Read: Labeling, Uncertainty Sampling

By Andrew Fairless on July 30, 2024May 24, 2024

https://code.groundlight.ai/python-sdk/blog/active-learning Reducing Data Labeling Costs with Uncertainty SamplingTed SandlerMay 15, 2024Senior Applied Scientist “With active learning, only informative images are prioritized for review, making it possible to label small aContinue readingWhat I Read: Labeling, Uncertainty Sampling

What I Read: Platonic Hypothesis

By Andrew Fairless on July 29, 2024May 24, 2024

https://phillipi.github.io/prh The Platonic Representation HypothesisMinyoung Huh, Brian Cheung, Tongzhou Wang, Phillip IsolaMITPosition Paper in ICML 2024 “Neural networks, trained with different objectives on different data and modalities, are converging toContinue readingWhat I Read: Platonic Hypothesis

Year: 2024