Base model: Qwen/Qwen3-30B-A3B
Architecture: Qwen3MoeForCausalLM
Transformers version in config: 4.57.1
Format: Hugging Face Transformers + Safetensors
Primary use: structured scoring, ranking, or classification prompts for economics-style research papers

Intended Use

The model is intended for research and internal experimentation on article-evaluation workflows, including tasks such as:

economics paper quality assessment
article-level prompt scoring
subject-specific taste benchmarking

It is not intended as a general-purpose factual assistant or as a substitute for expert peer review.

Files

This repository includes the files needed to load the model in Transformers:

model weights in sharded safetensors
tokenizer and chat template files
model config and generation config

The repository currently also contains training_args.bin from training export.

Example

python

from transformers import AutoModelForCausalLM, AutoTokenizer

model_id = "K1mG0ng/AI-taste-eco-30B-A3B"
tokenizer = AutoTokenizer.from_pretrained(model_id)
model = AutoModelForCausalLM.from_pretrained(model_id)

Notes

This is a fine-tuned research model.
Output quality depends heavily on prompt format and the label schema used at inference time.
Benchmark datasets and training code are maintained separately from this repository.