What are the hardware requirements for Sheared Pythia 160M?

Sheared Pythia 160M requires approximately 0.7 GB of VRAM and supports a context window of 2K tokens. Quantized variants may run on less VRAM; see the Quantized Models section on this page.

Who developed Sheared Pythia 160M and how large is it?

Sheared Pythia 160M is developed by princeton-nlp, a model with 160m parameters. The model is published as open weights on Hugging Face and indexed on LLM Explorer with full benchmark history.

How does Sheared Pythia 160M perform on standard benchmarks?

Sheared Pythia 160M has the following published scores: MMLU 26.65. Compare against reference models on this page or on the LLM Explorer leaderboards.

Sheared Pythia 160M by princeton-nlp — VRAM 0.7GB, 2K context

Name: Sheared Pythia 160M
Author: princeton-nlp

Sheared Pythia 160M is an open-source language model by princeton-nlp. Features: 160m LLM, VRAM: 0.7GB, Context: 2K, License: apache-2.0, LLM Explorer Score: 0.15, Arc: 22.4, HellaSwag: 32.1, MMLU: 26.7, GSM8K: 0.4.

Arxiv:2310.06694 Endpoints compatible Gpt neox Pytorch Region:us

Model Card on HF 🤗: https://huggingface.co/princeton-nlp/Sheared-Pythia-160m

Sheared Pythia 160M Parameters and Internals

Model Type

text generation

Training Details

Data Sources:

The Pile dataset

Data Volume:

0.4B tokens for pruning and 50B tokens for continued pre-training

Methodology:

Structured pruning

Model Architecture:

GPT-NeoX

LLM Name	Sheared Pythia 160M
Repository 🤗	https://huggingface.co/princeton-nlp/Sheared-Pythia-160m
Model Size	160m
Required VRAM	0.7 GB
Updated	2026-08-04
Maintainer	princeton-nlp
Model Type	gpt_neox
Model Files	0.7 GB
Model Architecture	GPTNeoXForCausalLM
License	apache-2.0
Context Length	2048
Model Max Length	2048
Transformers Version	4.28.1
Tokenizer Class	GPTNeoXTokenizer
Vocabulary Size	50304
Torch Data Type	float32

Rank the Sheared Pythia 160M capabilities

Have you tried this model? Rate its performance — this feedback helps the ML community find the right model for their needs.

Instruction Following and Task Automation
Factuality and Completeness of Knowledge
Censorship and Alignment
Data Analysis and Insight Generation
Text Generation
Text Summarization and Feature Extraction
Code Generation
Multi-Language Support and Translation

Best Alternatives to Sheared Pythia 160M

Best Alternatives	Context / RAM	Downloads	Likes
Pythia 160M C2s	8K / 0.6 GB	251	7
Violet 160M Chat	4K / 0.3 GB	13	0
Pythia 160M	2K / 0.4 GB	2294174	44
Pythia 160M Deduped	2K / 0.4 GB	133437	3
Pythia 160M Xsum Roya	2K / 0.6 GB	6	0
Pythia 160m Sft	2K / 0 GB	8	0
Pythia160m Sft Tldr	2K / 0.6 GB	175	0
Pythia 160M Dolphin Extended	2K / 0.3 GB	93	0
Pythia 160M Storytelling	2K / 0.3 GB	7	0
...hia 160M Deduped Step92k 193bt	2K / 0.4 GB	120	0

Note: green Score (e.g. "73.2") means that the model is better than princeton-nlp/Sheared-Pythia-160m.