www.interconnects.ai ↗︎
24 Pages Found
Synthetic data: Anthropic’s CAI, scaling, OpenAI’s Superalignment, tips, and open-source examples
www.interconnects.ai/p/llm-synthetic-data
Last glasped on Jan 1, 2024
3 users
Beyond human data: RLAIF needs a rebrand
www.interconnects.ai/p/beyond-human-data-rlaif
Last glasped on Aug 2, 2023
2 users
How RLHF actually works
www.interconnects.ai/p/how-rlhf-works
Last glasped on Jul 21, 2023
1 users
The Q* hypothesis: Tree-of-thoughts reasoning, process reward models, and supercharging synthetic data
www.interconnects.ai/p/q-star
Last glasped on Nov 29, 2023
1 users
A realistic path to robotic foundation models
www.interconnects.ai/p/robotic-foundation-models?utm_source=post-email-title&publication_id=48206&post_id=145280822&utm_campaign=email-post-title&isFreemail=true&r=18jahc&triedRedirect=true&utm_medium=email
Last glasped on Jun 5, 2024
1 users
Model merging lessons in The Waifu Research Department
www.interconnects.ai/p/model-merging
Last glasped on Apr 15, 2024
1 users
Futures of the data foundry business model
www.interconnects.ai/p/ai-data-foundry
Last glasped on Sep 12, 2024
1 users
The interface era of AI
www.interconnects.ai/p/the-interface-era-of-ai?utm_source=profile&utm_medium=reader2
Last glasped on Nov 29, 2023
1 users
State-space LLMs: Do we need Attention?
www.interconnects.ai/p/llms-beyond-attention
Last glasped on Feb 6, 2024
1 users
10 Sora and Gemini 1.5 follow-ups: code-base in context, deepfakes, pixel-peeping, inference costs…
www.interconnects.ai/p/sora-gemini-follow-up?utm_source=post-email-title&publication_id=48206&post_id=141829478&utm_campaign=email-post-title&isFreemail=false&r=3ax5p&utm_medium=email
Last glasped on Feb 20, 2024
1 users
Specifying objectives in RLHF
www.interconnects.ai/p/specifying-objectives-in-rlhf
Last glasped on Aug 3, 2023
1 users
LLAMA 2: an incredible open-source LLM
www.interconnects.ai/p/llama-2-from-meta
Last glasped on Jul 18, 2023
1 users
The interface era of AI
www.interconnects.ai/p/the-interface-era-of-ai
Last glasped on Nov 16, 2023
1 users
Synthetic data: Anthropic’s CAI, scaling, OpenAI’s Superalignment, tips, and open-source examples
www.interconnects.ai/p/llm-synthetic-data?utm_campaign=email-half-post&r=3ax5p&utm_source=substack&utm_medium=email
Last glasped on Nov 30, 2023
1 users
Mixtral: The best open model, MoE trade-offs, release lessons, Mistral raises $400mil, Google's loss, vibes vs marketing
www.interconnects.ai/p/mixtral
Last glasped on Dec 13, 2023
1 users
Why reward models are key for alignment
www.interconnects.ai/p/why-reward-models-matter?utm_source=post-email-title&publication_id=48206&post_id=141643312&utm_campaign=email-post-title&isFreemail=true&r=18jahc&utm_medium=email
Last glasped on Feb 14, 2024
1 users
Unfortunately, OpenAI and Google have moats
www.interconnects.ai/p/openai-google-llm-moats?utm_medium=ios
Last glasped on May 22, 2023
1 users
Interconnects | Nathan Lambert | Substack
www.interconnects.ai/p/gemma-google-ships-it?utm_source=post-email-title&publication_id=48206&post_id=141929268&utm_campaign=email-post-title&isFreemail=true&r=18jahc&open=false&utm_medium=email
Last glasped on Feb 22, 2024
1 users
Specifying hallucinations
www.interconnects.ai/p/specifying-hallucinations-llms
Last glasped on Jun 12, 2023
1 users
RLHF progress: Scaling DPO to 70B, DPO vs PPO update, Tülu 2, Zephyr-β, meaningful evaluation, data contamination
www.interconnects.ai/p/rlhf-progress-scaling-dpo-to-70b
Last glasped on Dec 3, 2023
1 users
Llama 3: Scaling open LLMs to AGI
www.interconnects.ai/p/llama-3-and-scaling-open-llms?utm_source=%2Fsearch%2FAGI&utm_medium=reader2
Last glasped on Apr 30, 2024
1 users
LLM products: measurement and manipulation
www.interconnects.ai/p/llm-products-measurement-manipulation?utm_source=profile&utm_medium=reader2
Last glasped on Sep 15, 2023
1 users
The DPO debate: Do we need RL for RLHF?
www.interconnects.ai/p/the-dpo-debate?utm_campaign=email-half-post&r=3ax5p&utm_source=substack&utm_medium=email
Last glasped on Dec 6, 2023
1 users
Local LLMs, some facts some fiction
www.interconnects.ai/p/local-llms?utm_campaign=email-half-post&r=3ax5p&utm_source=substack&utm_medium=email
Last glasped on Jan 25, 2024
1 users
Ready to highlight and find good content?
Glasp is a social web highlighter that people can highlight and organize quotes and thoughts from the web, and access other like-minded people’s learning.