Interconnects Audio Nathan Lambert

Many, many signs of life for preference fine-tuning beyond spoofing chat evaluation tools.This is AI generated audio with Python and 11Labs.Source code: https://github.com/natolambert/interconnects-toolsOriginal post: https://www.interconnects.ai/p/how-rlhf-works-2
00:00 How RLHF works, part 2: A thin line between useful and lobotomized04:27 The chattiness paradox08:09 The mechanism for making models chattier10:42 Next steps for RLHF research
Fig 1: https://huggingface.co/datasets/natolambert/interconnects-figures/resolve/main/rlhf/img_012.webpFig 2: https://huggingface.co/datasets/natolambert/interconnects-figures/resolve/main/rlhf/img_018.pngFig 3: https://huggingface.co/datasets/natolambert/interconnects-figures/resolve/main/rlhf/img_025.png

Models that seem totally out of scope from recent open LLMs give us a sneak peek of where the industry will be in 6 to 18 months.This is AI generated audio with Python and 11Labs.Source code: https://github.com/natolambert/interconnects-toolsOriginal post: https://www.interconnects.ai/p/phi-3-and-arctic-llms
0:00 Phi 3 and Arctic: Outlier LMs are hints1:01 Arctic & open mixture of expert trends6:10 Phi 3, synthetic data, and small models
Fig 1: https://huggingface.co/datasets/natolambert/interconnects-figures/resolve/main/phi3/img_004.pngFig 2: https://huggingface.co/datasets/natolambert/interconnects-figures/resolve/main/phi3/img_008.pngFig 3: https://huggingface.co/datasets/natolambert/interconnects-figures/resolve/main/phi3/img_018.png

Certain definitions of AGI are backing people into a pseudo-religious corner.This is AI generated audio with Python and 11Labs.Source code: https://github.com/natolambert/interconnects-toolsOriginal post: https://www.interconnects.ai/p/agi-is-what-you-want-it-to-be
00:00 AGI is what you want it to be04:01 RL still rules the AGI discourse05:43 Modern AGI tests07:37 Agency and shifting goalposts
Fig 1: https://huggingface.co/datasets/natolambert/interconnects-figures/resolve/main/agi/img_018.pngFig 2: https://huggingface.co/datasets/natolambert/interconnects-figures/resolve/main/agi/img_020.png

Meta shows that scaling won't be a limit for open LLM players in the near future.This is AI generated audio with Python and 11Labs.Source code: https://github.com/natolambert/interconnects-toolsOriginal post: https://www.interconnects.ai/p/llama-3-and-scaling-open-llms
00:00 Llama 3; scaling open LLMs to AGI01:44 Pretraining, data, and basic evals06:06 Alignment and human evaluations10:08 Chatting with Meta AI and Llama 3 70B Instruct11:55 Same Llama license (mostly)12:52 The healthy open LLM ecosystem
Fig 1: https://huggingface.co/datasets/natolambert/interconnects-figures/resolve/main/llama3/img_011.jpegFig 2: https://huggingface.co/datasets/natolambert/interconnects-figures/resolve/main/llama3/img_013.pngFig 3: https://huggingface.co/datasets/natolambert/interconnects-figures/resolve/main/llama3/img_015.pngFig 4: https://huggingface.co/datasets/natolambert/interconnects-figures/resolve/main/llama3/img_020.pngFig 5: https://huggingface.co/datasets/natolambert/interconnects-figures/resolve/main/llama3/img_036.pngFig 6: https://huggingface.co/datasets/natolambert/interconnects-figures/resolve/main/llama3/img_040.pngFig 7: https://huggingface.co/datasets/natolambert/interconnects-figures/resolve/main/llama3/img_046.jpegFig 8: https://huggingface.co/datasets/natolambert/interconnects-figures/resolve/main/llama3/img_061.pngFig 9: https://huggingface.co/datasets/natolambert/interconnects-figures/resolve/main/llama3/img_063.webpFig 10: https://huggingface.co/datasets/natolambert/interconnects-figures/resolve/main/llama3/img_066.pngFig 11: https://huggingface.co/datasets/natolambert/interconnects-figures/resolve/main/llama3/img_068.jpeg

Integrating some non computing science into reinforcement learning from human feedback can give us the models we want.This is AI generated audio with Python and 11Labs.Source code: https://github.com/natolambert/interconnects-toolsOriginal post: https://www.interconnects.ai/p/reinventing-llm-alignment
0:00 Stop "reinventing" everything to "solve" AI alignment2:19 Social Choice for AI Alignment: Dealing with Diverse Human Feedback7:03 OLMo 1.7 7B: A truly open model with actually good benchmarks
Fig 1: https://huggingface.co/datasets/natolambert/interconnects-figures/resolve/main/reinvention/img_013.pngFig 2: https://huggingface.co/datasets/natolambert/interconnects-figures/resolve/main/reinvention/img_015.pngFig 3: https://huggingface.co/datasets/natolambert/interconnects-figures/resolve/main/reinvention/img_018.pngFig 4: https://huggingface.co/datasets/natolambert/interconnects-figures/resolve/main/reinvention/img_024.pngFig 5: https://huggingface.co/datasets/natolambert/interconnects-figures/resolve/main/reinvention/img_027.png

Modeling the compute versus performance tradeoff of many open LLMs.This is AI generated audio with Python and 11Labs.Source code: https://github.com/natolambert/interconnects-toolsOriginal post: https://www.interconnects.ai/p/compute-efficient-open-llms
0:00 The end of the "best open LLM"3:05 Compute efficient open LLMs
Fig 1: https://huggingface.co/datasets/natolambert/interconnects-figures/resolve/main/scaling/img_004.jpegFig 2: https://huggingface.co/datasets/natolambert/interconnects-figures/resolve/main/scaling/img_009.pngFig 3: https://huggingface.co/datasets/natolambert/interconnects-figures/resolve/main/scaling/img_014.pngFig 4: https://huggingface.co/datasets/natolambert/interconnects-figures/resolve/main/scaling/img_016.pngFig 5: https://huggingface.co/datasets/natolambert/interconnects-figures/resolve/main/scaling/img_018.pngFig 6: https://huggingface.co/datasets/natolambert/interconnects-figures/resolve/main/scaling/img_020.pngFig 7: https://huggingface.co/datasets/natolambert/interconnects-figures/resolve/main/scaling/img_022.pngFig 8: https://huggingface.co/datasets/natolambert/interconnects-figures/resolve/main/scaling/img_024.pngFig 9: https://huggingface.co/datasets/natolambert/interconnects-figures/resolve/main/scaling/img_028.png