NoPos
No estimate
No hardware requirements for this model
The weights for this model have not been published, so it cannot be downloaded or run on your own hardware at any size. It is reachable only through its provider, and no graphics card changes that.
On record
Full specification
Everything on record for this model. Most of it describes how it was trained rather than how it runs — useful context for judging how much work went into it, and how it compares with models built at a different scale.
Origin
Who built this model, where, and when it was published.
- Organisation
- Tel Aviv University,University of Washington,Intel Labs,Meta AI
- Organisation type
- Academia,Academia,Industry,Industry
- Country
- Israel, United States of America
- Published
- 30 March 2022
- Authors
- Adi Haviv, Ori Ram, Ofir Press, Peter Izsak, Omer Levy
What it does
The problem areas the model was built for. A model can carry several of each.
- Domain
- Language
- Task
- Language modeling/generation
Size
How large the model is and how much data it was trained on. Parameters are the figure that decides whether it fits on a given graphics card.
- Parameters
- 1.3B
- Training data
- 21,000,000,000 tokens
- Epochs
- 0.12
- Batch size
- 256,000
1.3B "The baseline model in this setting follows the 1.3B parameter architecture of Brown et al. (2020), also known as GPT-3 XL: 24 transformer layers with 2048 model dimensions, 8192 feed-forward dimensions, and 32 attention heads."
Table 5
Training compute
The arithmetic performed to train the model, measured in floating-point operations. It is a measure of what the training run cost, not of how fast the finished model answers you.
- Training compute
- 2.1 × 10¹⁹ FLOP
- How it was established
- Operation counting
6 FLOP / token / parameter * 1.3 * 10^9 parameters * 256 000 tokens per batch [table 5] * 10500 steps [table 5] = 2.09664e+19 FLOP
Availability
Whether you can obtain the model and run it on your own hardware, which is what decides if any of the graphics-card figures on this page apply.
- Weights
- Closed — provider access only
- Model access
- Unreleased
- Training code
- Unreleased
How it is classified
Labels the source dataset applies when tracking notable models, and how confident it is in the entry.
- Record confidence
- Confident
- Citations
- 185
- Benchmark data
- NoPos
Sources
Where this record came from and when it was last checked.
- Reference
- Transformer Language Models without Positional Encodings Still Learn Positional Information
- Last updated
- 25 May 2026
What the numbers mean
Where it came from
NoPos was published by Tel Aviv University,University of Washington,Intel Labs,Meta AI, in Israel, in March 2022. It comes out of academia,Academia,Industry,Industry.
It works in Language, and is recorded as doing language modeling/generation.
Because the weights are not available, none of the hardware figures elsewhere on this site apply to it.
What went into building it
The training run consumed about 2.1 × 10¹⁹ FLOP. That figure describes the cost of creating it and has no bearing on how quickly it generates text.
The training set ran to roughly 21,000,000,000 tokens.
Answers
NoPos — common questions
What GPU do I need to run NoPos?
None. NoPos is a closed model — its weights were never published, so it cannot be downloaded or run on your own hardware at any price. It is reachable only through its provider.
Is NoPos open source?
No. NoPos has not had its weights published, so it exists only as a service controlled by its owner.
How many parameters does NoPos have?
NoPos has 1.3B parameters. 1.3B "The baseline model in this setting follows the 1.3B parameter architecture of Brown et al. (2020), also known as GPT-3 XL: 24 transformer layers with 2048 model dimensions, 8192 feed-forward dimensions, and 32 attention heads.". That figure is the total, and it is what decides how much memory the model needs — roughly half a gigabyte per billion at the compression most people use.
Who created NoPos?
NoPos was published by Tel Aviv University,University of Washington,Intel Labs,Meta AI, based in Israel, categorised as academia,Academia,Industry,Industry.
When was NoPos released?
NoPos was published in March 2022. Capability per parameter has improved considerably since, so a newer model of the same size is often the better use of the same hardware.
What is NoPos used for?
NoPos works in Language, and is recorded as handling language modeling/generation. These are the areas it was designed around; they describe intent rather than a hard boundary.
How much compute was used to train NoPos?
Around 2.1 × 10¹⁹ FLOP. That measures what producing the model cost and says nothing about how quickly it answers once trained — inference speed comes from memory bandwidth, not from the training budget.
The other direction
Looking at it from the other side?
This page starts from the model. If you already own a card and want to know everything it will run, start from the hardware instead.