aaravshirpurkar
10 hours ago
Fine-tuned the Qwen3:4B model on Advaita Vedanta text, mainly Ashtavakra Gita, Mandukya Upanishad and a few other primary Advaita Vedanta texts. Made my own dataset from these sources and then fine-tuned it on Kaggle free T4 GPU.
Did this experiment too see if the model can recognize the patterns of Advaita Vedanta texts and topics like consciousness, awareness, reality etc. and can it mimic the same patterns or pretend it's conscious.. did not get that answer yet but it had some interesting results
Model+results: https://huggingface.co/aaravshirpurkar/turiya-model
Dataset: https://huggingface.co/datasets/aaravshirpurkar/turiya_datas...