Who developed Instruct GPT J and how large is it?

Instruct GPT J is developed by crumb. The model is published as open weights on Hugging Face and indexed on LLM Explorer with full benchmark history.

Where can I download or evaluate Instruct GPT J?

Instruct GPT J is hosted on Hugging Face and linked from this page. LLM Explorer also lists quantized variants and similar alternatives if available.

Name: Instruct GPT J
Author: crumb

Instruct GPT J is an open-source language model by crumb. Features: LLM, Instruction-Based, LLM Explorer Score: 0.06.

Model Type

text generation, instruction-following

Use Cases

Areas:

Research, Instruction-following applications

Applications:

Instruction-based text generation tasks

Primary Use Cases:

Generating responses to instructional prompts

Additional Notes

This model is a personal project and not officially affiliated with EleutherAI.

Supported Languages

en (high)

Training Details

Data Sources:

tatsu-lab/alpaca

Methodology:

Finetuned using LoRA (Low-Rank Adaptation)

Model Architecture:

Based on GPT-J architecture

Input Output

Input Format:

Instruction and optionally paired input context.

Accepted Modalities:

text

Output Format:

Generated text response to the instruction.

LLM Name	Instruct GPT J
Repository 🤗	https://huggingface.co/crumb/Instruct-GPT-J
Required VRAM	0 GB
Updated	2026-07-18
Maintainer	crumb
Instruction-Based	Yes
Model Files	0.0 GB
Supported Languages	en
Model Architecture	AutoModel
Is Biased	none
PEFT Type	LORA
LoRA Model	Yes
PEFT Target Modules	q_proj\|v_proj\|k_proj\|out_proj
LoRA Alpha	32
LoRA Dropout	0.05
R Param	8

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	Context / RAM	Downloads	Likes
...l 8x7B Instruct V0.1 Llamafile	0K / GB	7076	18
Autotrain Umberto Proclama	0K / 0.9 GB	5	0
Mamba Python	0K / 2 GB	13	0
...l 8x7B Instruct V0.1 Llamafile	0K / GB	3180	18
...hi 3 Mini 4K Instruct Ct2 Int8	0K / 3.8 GB	8	1
CSUMLM	0K / GB	13	1
...hin 2.5 Mixtral 8x7b Llamafile	0K / GB	117	4
Vigogne Bloom 7b1 Instruct	0K / 0.1 GB	0	4
...a Instruction Fine Tune French	0K / 0 GB	0	4
MiniMaid L3	0K / 0 GB	9	2

Note: green Score (e.g. "73.2") means that the model is better than crumb/Instruct-GPT-J.