I’m just guessing this is the reason but the modal answer from all human statements about the count of R’s in “strawberry” might actually be 2. So it does it because we do.
RESEARCH
-
Suspicion about poor generalization on very hard words like strawberry
By
–
I suspect this might not generalize to very hard words like “strrrraaawwberrrrrrry” etc.
-
Counting R’s in ‘strawberry’ with chain-of-thought as LLM psychology
By
–
There’s no reason to do this except as an exercise in LLM psychology, but here’s how you could count how many R’s are in “strawberry” using chain-of-thought. (Normally, letter- or count-based tasks like this would be handled outside the LLM, by having it call an API or REPL.)
-
From Discoveries to Tactics to Strategic AI Implementation
By
–
let discoveries become tactics and tactics become strategy
-
Patterns for Building LLM-based Systems and Products
By
–
Patterns for Building LLM-based Systems & Products https://
bit.ly/45CdmrF
#AI #MachineLearning #DeepLearning #LLMs #DataScience -
Sparse-IFT Transformations Presented at ICML 2024
By
–
Highlights from #ICML 2024 ICML 2024 has been incredible! We've enjoyed insightful discussions and the opportunity to connect with AI researchers and practitioners. We're proud to present the following papers at ICML 2024: "Sparse-IFT: Sparse Iso-FLOP Transformations
-
Insufficient Training Data for Spelling Tasks in LLMs
By
–
I think there’s not enough training data naturally on the internet of spelling tasks compared to the difficulty of the task for the LLM, due to how text is chopped up into sequences of text chunks (tokens), all of which are unique / distinct. I have a whole video on Tokenization.
-
Neural Scaling Laws Workshop Highlights from ICML 2024
By
–
Highlights from #ICML 2024! We had an incredible time at the 7th Workshop on Neural Scaling Laws: Scaling, Transfer & Multilingual Models, held earlier this week at ICML in Vienna. This joint event, co-organized with Irina Rish and CERC-AAI, brought together brilliant
-
DeepMind and OpenAI: Comparing AI Research and Education Impact
By
–
Gary, you’re right that Google DeepMind has been just as amazing as OpenAI. Moreover, I admire DeepMind’s contributions to AI education in Africa and LatAm over the years. I truly love DeepMind. However, I did not mean to invite comparisons. What I really intended to provide as
-
Learning Orthogonalized Matrices vs Computing Refusal Directions
By
–
I'd say it's different because you learn the orthogonalized matrix with OFT while you compute (you don't learn) your refusal direction here