Pinned Loading

1. BlicketTest_CausalReasoning BlicketTest_CausalReasoning Public

   Built a multi-turn RL training environment that can be used to benchmark and improve LLM causal reasoning based on the Blicket Detector Test from developmental psychology, motivated by published re…

   Python

2. CausalReasoningEnv CausalReasoningEnv Public

   Developed an environment to benchmark and train an LLM's ability to reason and identify causal effects on Directed Acyclic Graphs that represent structural causal models

   Python

3. Unsupervised-Elicitation Unsupervised-Elicitation Public

   Forked from Jiaxin-Wen/Unsupervised-Elicitation

   Implementation of Internal Coherence Maximization on TruthfulQA -- results suggest unsupervised LLM alignment may be possible

   Python

4. Persona-Elicitation Persona-Elicitation Public

   Implementation of Internal Coherence Maximization for Persona Elicitation on the GlobalOpinionQA dataset

   Python

5. Reasoning-Theater Reasoning-Theater Public

   Forked from AskSid/disentangling-computation-from-cot

   Python

6. ReinforcementLearning ReinforcementLearning Public

   This repo contains scripts and notebooks in which I explore and implement common RL algorithms on agents operating in Gym envs

   Jupyter Notebook