Projects

Research

Alignment Pretraining: AI Discourse Causes Self-Fulfilling (Mis)alignment

Cameron Tice, Puria Radmard, Samuel Ratnam, Andy Kim, David Africa, Kyle O'Brien

How discussions of AI in pretraining data shape model behavior through self-fulfilling mechanisms.

arxiv →

Why Jailbreaks Look Like Hypnosis: Parallels Between Human and LLM Cognition

Samuel Ratnam

Theoretical work on the parallels between human and LLM cognition.

pdf →

Objective Matters: Fine-Tuning Objectives Shape Safety, Robustness, and Persona Drift

Daniel Vennemeyer, Punya Syon Pandey, Phan Anh Duong, Michael Umeokoli, Samuel Ratnam

How different fine-tuning objectives impact language model safety and adversarial robustness at scale.

arxiv →

Code

OpenRecommender

A locally-running, AI-powered content recommendation system that prioritizes user transparency and control.

github →

Idealists Collective

Website for a collective of philosophers, artists, and technologists with git-tracked essay revisions.

github →

Non-Linear Writing Interface

A keyboard-native writing application for crafting branching narratives using DAG visualization.

github →

JobGPT

AI-powered interview practice with GPT-generated questions and automated feedback. Winner of Bincom Hackathon 2023.

github →

Poetry Copilot

AI writing assistant for poetry, providing inspiration, intelligent thesaurus, and editing suggestions.

github →

Iris

Personal AI assistant Discord bot integrating Google Calendar, Gmail, Drive, and Todoist.

github →

Favour Auction

A point-based favour trading platform where users submit favours and bid in auctions.

github →