LLM Parameter Sizes

Comments discuss the parameter counts of large language models, debating what constitutes a 'large' or 'small' model, comparisons to benchmarks like GPT-3, and topics like quantization, inference efficiency, and scaling.

➡️ Stable 0.6x AI & Machine Learning
5,168
Comments
16
Years Active
5
Top Authors
#6885
Topic ID

Activity Over Time

2008
1
2011
1
2013
1
2014
2
2015
4
2016
13
2017
16
2018
17
2019
73
2020
169
2021
164
2022
274
2023
1,894
2024
1,187
2025
1,296
2026
56

Keywords

RAM GPT2 LM M40 MacBook BLOOM RL RPI github.com GPT parameters model gpt models vram language model training size 5b ram

Sample Comments

amznbyebyebye Dec 8, 2021 View on HN

Pffft only 280B parameters? Give me a break

m3kw9 Mar 29, 2023 View on HN

How does training on just 800k pieces of data need 7b parameters?

stavros Jan 6, 2024 View on HN

Is it really close to GPT-3.5 at 2.7B?

dharma1 Mar 1, 2023 View on HN

Reckon they will (if not already) use 4bit or 8bit precision and may not need 175b params

tama_sala Jul 16, 2024 View on HN

The embedding size is only 8k so while the parameters are 70B. So it's a huge difference

SemanticStrengh May 18, 2022 View on HN

28.7 million parameter is nothing for inference

liminal Dec 16, 2024 View on HN

Is 14B parameters still considered small?

euclaise Jan 20, 2023 View on HN

Assuming you're referring to the largest model - BLOOM is huge, this is not, so presumably much worse

croes May 24, 2025 View on HN

How can a Large Language Model be a small language model?

rubslopes Mar 16, 2025 View on HN

Is a tiny large language model equivalent to a normal sized one?