Recent Posts

• Gemini Flash 2.0 Thinking Experimental
  by convexstrictlyon 12/19/2024, 4:44:19 PM with 3 comments
• What Questions Are in the Chinese College Entrance Exam?
  by convexstrictlyon 6/14/2024, 11:46:26 AM with 0 comments
• Generative AI Is Not Going to Build Your Engineering Team for You
  by convexstrictlyon 6/14/2024, 11:42:40 AM with 0 comments
• Building GPT2o – Part 1: Audio
  by convexstrictlyon 6/14/2024, 11:26:42 AM with 1 comment
• The Geometry of Categorical and Hierarchical Concepts in Large Language Models
  by convexstrictlyon 6/8/2024, 9:49:00 PM with 0 comments
• OpenAI says it has begun training a new flagship A.I. model
  by convexstrictlyon 5/28/2024, 10:47:25 AM with 0 comments
• California residents: call your legislators about AI bill SB 1047
  by convexstrictlyon 5/20/2024, 11:38:04 PM with 5 comments
• LISA: Layerwise Importance Sampling for Memory-Efficient LLM Fine-Tuning
  by convexstrictlyon 3/27/2024, 5:17:15 PM with 1 comment
• NTIA AI Open Model Weights RFC
  by convexstrictlyon 3/27/2024, 12:29:14 AM with 1 comment
• Mechanics of Next Token Prediction with Self-Attention
  by convexstrictlyon 3/19/2024, 3:11:11 PM with 0 comments
• Dive Deeper into Yi-9B
  by convexstrictlyon 3/18/2024, 9:53:15 PM with 0 comments
• You can now train a 70B language model at home
  by convexstrictlyon 3/6/2024, 6:51:47 PM with 1 comment
• Shape Suffixes – Good Coding Style
  by convexstrictlyon 2/28/2024, 1:50:27 PM with 0 comments
• Star Trek prompt optimal for grade school math on Llama-70B
  by convexstrictlyon 2/26/2024, 3:28:33 PM with 1 comment
• (US Dept of Commerce) NTIA Solicits Comments on Open-Weight AI Models
  by convexstrictlyon 2/23/2024, 9:46:24 PM with 0 comments
• BitDelta: Your Fine-Tune May Only Be Worth One Bit
  by convexstrictlyon 2/16/2024, 4:57:31 PM with 2 comments
• Time is encoded in the weights of finetuned language models
  by convexstrictlyon 12/24/2023, 9:53:03 PM with 10 comments
• Zoology 1: Measuring and Improving Recall in Efficient Language Models
  by convexstrictlyon 12/22/2023, 4:00:41 PM with 1 comment
• TinyGSM: Achieving >80% on GSM8k with small language models
  by convexstrictlyon 12/15/2023, 2:44:33 AM with 1 comment
• Androids built to meet the labor demands
  by convexstrictlyon 12/1/2023, 4:16:36 PM with 1 comment
• Sam Altman likely to start company with researchers from OpenAI: Bloomberg
  by convexstrictlyon 11/18/2023, 7:15:41 PM with 2 comments