whats your name and how many parameters you got
- User
whats your name and how many parameters you got
Assistant
I am a large language model, trained by Google.
The exact number of parameters I have isn't publicly disclosed. Large language models, especially recent ones developed by companies like Google, often have parameter counts kept confidential.
However, Google has developed several large language models over time with varying parameter counts:
- Gemini: This is Google's current family of flagship models (Ultra, Pro, Nano). They are multimodal (handling text, code, audio, image, video) and state-of-the-art, but Google hasn't released specific parameter counts for them.
- PaLM (Pathways Language Model): Released in 2022, this model has 540 billion parameters[1][2].
- Gemma: These are open-source models trained using similar resources as Gemini[3]. Gemma 2, released in 2024, comes in 9 billion and 27 billion parameter sizes[3].
- LaMDA (Language Model for Dialogue Applications): Announced in 2021, parameter counts weren't the focus of its release.
- BERT: Introduced in 2018, BERT had 340 million parameters.
- Switch Transformer: A 2021 research paper discussed a model using this technique with up to 1.6 trillion parameters, though this was more experimental.
It's important to note that parameter count isn't the only measure of a model's capability[4]. Factors like architecture, training data, and efficiency also play significant roles.
Other models, same or similar prompt
Responses were produced in different, uncontrolled contexts (system prompts, settings, prior turns). This is not a benchmark.