2 3 2

Mac Szankin

macsz

https://szank.in/

AI & ML interests

LLMs, Autonomous spaceships and GenAI at SiMa.ai. After-hours Projects: Processing Thermal Imagery.

Recent Activity

upvoted a collection about 16 hours ago

Open LLM Leaderboard best models ❤️‍🔥

liked a model 20 days ago

sentence-transformers/all-MiniLM-L12-v2

replied to fdaudens's post 7 months ago

🤖💡Just tried out @m-ric 's new LLaMA-3.1 70B agent for data analysis. Impressive stuff. 🚢📊 Fed it the Titanic passenger dataset with minimal instructions. The agent autonomously dug in, tested hypotheses, and reached some intriguing conclusions: "Lower class passengers less likely to survive, slight negative correlation with age, and positive correlation between fare price and survival." 📈It even generated charts to visualize the findings! 🧠💼 Great potential for business intelligence, research, and decision-making when we can just upload datasets and let AI agents loose on them. 👉 Check it out: https://huggingface.co/spaces/m-ric/agent-data-analyst 🤔 Any particular use cases you're excited about? #AIinDataAnalysis #MachineLearning #DataScience

View all activity

Organizations

macsz's activity

upvoted a collection about 16 hours ago

Open LLM Leaderboard best models ❤️‍🔥

Collection

A daily uploaded list of models with best evaluations on the LLM leaderboard: • 64 items • Updated about 2 hours ago • 546

liked a model 20 days ago

sentence-transformers/all-MiniLM-L12-v2

replied to fdaudens's post 7 months ago

Missed it! Thanks! This is 🔥

New activity in meta-llama/Llama-3.1-8B-Instruct 7 months ago

ValueError: `rope_scaling` must be a dictionary with two fields

#15 opened 7 months ago by

jsemrau

upvoted a paper 7 months ago

ThermalNeRF: Thermal Radiance Fields

Paper • 2407.15337 • Published Jul 22, 2024 • 5

reacted to morgan's post with 👍 7 months ago

Post

1304

Llama 3.1 405B Instruct beats GPT-4o on MixEval-Hard

Just ran MixEval for 405B, Sonnet-3.5 and 4o, with 405B landing right between the other two at 66.19

The GPT-4o result of 64.7 replicated locally but Sonnet-3.5 actually scored 70.25/69.45 in my replications 🤔 Still well ahead of the other 2 though.

Sammple of 1 of the eval calls here: https://wandb.ai/morgan/MixEval/weave/calls/07b05ae2-2ef5-4525-98a6-c59963b76fe1

Quick auto-logging tracing for openai-compatible clients and many more here: https://wandb.github.io/weave/quickstart/