Intro to RLHF

DPO v/s PPO v/s GRPO In order to understand what these fancy “alignement” algorithms mean, let’s go back to the basics first - RLHF. There are many applications such as writing stories where you want creativity, pieces of informative text which should be truthful, or code snippets that we want to be executable. Writing a loss function to captu... Read more 16 Feb 2025 - 8 minute read

Change a Bit to save Bytes: Compression for Floating Point Time-Series Data

Reading this paper for GSOC 2025 - CERN. This paper focuses on data manipulations prior to using a compression algorithm. The highlight are two data preprocessing algorithms - Addition and Multiplication. Employing these algorithms improve the compressor effectiveness by increasing the number of bit values shared in the dataset to reduce its en... Read more 16 Feb 2025 - 5 minute read

B2B

I have always been interested in the functioning and organization of brain. Some people have been saying the age of B2B SAAS is over, it’s the age of Agent as a service. But we are at the dawn of an era where we are going to experince a revolutionary shift in the way we consume and transmit information. To back this stupendous claim let me attac... Read more 11 Feb 2025 - 2 minute read

Hello World

Hello Everyone! Samkit this side. Going public today, embarking on my journey to the glory. So much to do, so little time. Read more 10 Feb 2025 - less than 1 minute read

Older Newer