Ansh Radhakrishnan's repositories
trl_custom
Applying Reinforcement Learning from Human Feedback to language models to teach them to write short story responses to writing prompts.
Dalle-Mini-RL
Fine-tuning Dalle-Mini with RL to not produce NSFW images
dalle-mini
DALL·E Mini - Generate images from a text prompt
deep_learning_curriculum
Language model alignment-focused deep learning curriculum
elk
Keeping language models honest by directly eliciting knowledge encoded in their activations. Building on "Discovering latent knowledge in language models without supervision" (Burns et al. 2022)
equinox
Callable PyTrees and filtered transforms => neural networks in JAX. https://docs.kidger.site/equinox/
fancy_einsum
Einsum with einops style variable names
flax_minimal_gpt
This is a minimal implementation of a GPT style transformer model in FLAX, mostly done for learning purposes.
littlebookofsemaphores
Python answers to Puzzles in The Little Book of Semaphores
ml-interviews-book-answers
Answers to https://huyenchip.com/ml-interviews-book/
mlab2_pre_exercises
Pre-course exercises for the August 2022 MLAB cohort.
Module-0
Module 0 - Fundamentals
Module-1
Module 1 - Autodifferentiation
Module-2
Module 2 - Tensors
Module-3
Module 3 - Efficiency
numpy-100
100 numpy exercises (with solutions)
pytorch
Tensors and Dynamic neural networks in Python with strong GPU acceleration
Sorting-Transformer-Interp
A mechanistic interpretability project meant to analyze how a simple transformer learns to sort a sequence of 10 digits.