Mirror of your own self – on AI sycophancy and getting what we asked forsubstack.com·3 pts·yenniejun111·0
Visual Hallucinations in Psychedelics vs. Schizophreniapsychedelirium.substack.com·2 pts·yenniejun111·0
A General Theoretical Paradigm to Understand Learning from Human Preferencesarxiv.org·2 pts·yenniejun111·0
Is the hallucinogenic "trip" needed for therapeutic benefits from psychedelics?psychedelirium.substack.com·5 pts·yenniejun111·5
More Meaningful Evaluations for Values and Opinions in Large Language Modelsarxiv.org·2 pts·yenniejun111·0
Google DeepMind trained a robot to beat humans at table tennistechnologyreview.com·1 pts·yenniejun111·0
A Large-Scale Multimodal Dataset with Multigranular Annotations for Medicinehuggingface.co·2 pts·yenniejun111·0
From GPT-2 to GPT-4: How LLMs understand different occupations over timeartfish.ai·1 pts·yenniejun111·1