Will a Mamba 7b model trained on 2 trillion tokens outperform Llama2-13B
21
1kṀ738Jul 1
66%
chance
1D
1W
1M
ALL
Question will resolve positive if someone trains a Mamba (https://twitter.com/tri_dao/status/1731728602230890895) language model with <=7.5billion parameters on <=2 trillion tokens that outperforms Llama2-13B on the huggingface open llm leaderboard (https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard)
This question is managed and resolved by Manifold.
Get
1,000 to start trading!
1,000Related questions
Related questions
Will anyone train a TokenFormer model at scale before 2026?
25% chance
What will be true of the first model to cross 1400 on lmarena.ai?
Will Llama 4 be the best LLM in the chatbot arena?
10% chance
Will a flagship (>60T training bytes) open-weights LLM from Meta which doesn't use a tokenizer be released in 2025?
29% chance
Will the next major LLM by OpenAI use a new tokenizer?
77% chance
Will a single model running on a single consumer GPU (<1.5k 2020 USD) outperform GPT-3 175B on all benchmarks in the original paper by 2025?
86% chance
Before 2028, will any AI model achieve the same or greater benchmarks as o3 high with <= 1 million tokens per question?
87% chance
How many active parameters will the largest Llama 3 have?
77% chance
Will Llama 3-multimodal be natively mixed-multimodal? (VQ-VAE+next token prediction)
50% chance
Will the next LLM released by OpenAI be worse than GPT-4 at MMLU?
16% chance