PinnedI truly appreciate your kind words, It’s an honor to be acknowledged by the author, and I’m really…Jan 31Jan 31
PinnedThanks for the appreciation, Its surreal for me to get acknowledged from the author itself.Feb 1, 2024Feb 1, 2024
Papers Explained 340: CHASECHallenging AI with Synthetic Evaluations (CHASE) is a unified framework to synthetically generate challenging problems using LLMs without…Mar 28Mar 28
Papers Explained 339: Code Guided Synthetic data generation system (CoSyn)Code Guided Synthetic data generation system (CoSyn) is a framework that leverages the coding capabilities of text-only LLMs to…Mar 27Mar 27
Papers Explained 338: Large-Scale Data Selection for Instruction TuningThis work presents a systematic study of how well data selection methods scale, It finds that:Mar 26Mar 26
Papers Explained 337: Logic-RLThis study explores the potential of rule-based reinforcement learning (RL) in large reasoning models. Synthetic logic puzzles are used as…Mar 25Mar 25
Papers Explained 336: Rethinking Compute-Optimal Test-Time ScalingThis paper focuses on two core questions:Mar 24Mar 24
Papers Explained 335: Transformers without NormalizationThis work demonstrates that Transformers without normalization can achieve the same or better performance using a remarkably simple…Mar 211Mar 211
Papers Explained 334: Kimi k1.5Kimi k1.5 multi-modal LLM trained with RL, including its RL training techniques, multi-modal data recipes, and infrastructure optimization…Mar 20Mar 20
Papers Explained 333: SmolDoclingSmolDocling is a 256M parameter vision-language model Based on Hugging Face’s SmolVLM designed for end-to-end document conversion. It…Mar 19Mar 19