legesher/language-decoded-lora

Language Decoded LoRA

QLoRA adapters fine-tuned on multilingual code conditions for the Language Decoded project (part of Cohere's Tiny Aya Expedition).

Research Question

Does fine-tuning on non-English code improve multilingual reasoning — and is the benefit language-dependent or structure-dependent?

Base Model

All adapters are trained on CohereLabs/tiny-aya-base (3.35B parameters).

Model Structure

This repo is the canonical hub for all Language Decoded LoRA adapters, organized by experimental condition:

Subdirectory	Condition	Training Data
condition-1-en-32k/	Condition 1	English Python from The Stack Dedup (full 32k corpus)
condition-1-en-5k/	Condition 1	English Python from The Stack Dedup (5k subset)
condition-2-zh-5k/	Condition 2	Chinese keyword-swapped Python (Legesher-transpiled)
condition-2-es-5k/	Condition 2	Spanish keyword-swapped Python (Legesher-transpiled)
condition-2-ur-5k/	Condition 2	Urdu keyword-swapped Python (Legesher-transpiled)
condition-3-zh-5k/	Condition 3	Transpiled + native Chinese code (blended)

The Experimental Ladder

• Baseline --> 1: Does code help at all?
• 1 --> 2: Does the language of keywords matter?
• 2 --> 3: Does diversity of native-language sources add value beyond keyword swap?
• 3 --> 4: Does code written in the cultural context of a language carry unique signal?

Usage

from transformers import AutoModelForCausalLM, AutoTokenizer
from peft import PeftModel

# Load base model
base_model = AutoModelForCausalLM.from_pretrained("CohereLabs/tiny-aya-base")
tokenizer = AutoTokenizer.from_pretrained("CohereLabs/tiny-aya-base")

# Load a LoRA adapter (e.g., Condition 1 — English code)
model = PeftModel.from_pretrained(base_model, "legesher/language-decoded-lora", subfolder="condition-1-en-5k")

# Load a language-specific adapter (e.g., Condition 2 — Chinese keyword-swapped)
model = PeftModel.from_pretrained(base_model, "legesher/language-decoded-lora", subfolder="condition-2-zh-5k")

Training Details

Parameter	Value
Base model	CohereLabs/tiny-aya-base (3.35B params)
Method	QLoRA 4-bit (NF4), ~5.4GB VRAM
Hardware	Kaggle T4 (16GB)
Tokenizer	CohereLabs/tiny-aya-base
Transpilation tool	Legesher v0.7.3
Training data	legesher/language-decoded-data

QLoRA Hyperparameters

Parameter	Value
LoRA rank (r)	16
LoRA alpha	32
LoRA dropout	0.0
Target modules	q_proj, k_proj, v_proj, o_proj, up_proj, down_proj, gate_proj
Bias	none
Task type	CAUSAL_LM
PEFT version	0.18.1
Quantization	NF4 (4-bit) via Unsloth

Evaluation

Models are evaluated on multilingual reasoning benchmarks with dual prompts (English + language-specific):

Benchmark	What it measures	Examples per language
MGSM	Math reasoning	250 (full set)
X-CSQA	Commonsense reasoning	~1,000 (full set)
XNLI	Natural language inference	~5,000 (full set)

Results will be added as evaluation completes.

Limitations

• Single base model: All adapters are trained on CohereLabs/tiny-aya-base (3.35B params). Results may not generalize to larger or architecturally different models.
• Limited training data: Each condition uses a 5k-file subset for QLoRA fine-tuning, constrained by Kaggle T4 hardware limits.
• Evaluation scope: Currently evaluated on 3 benchmarks (MGSM, X-CSQA, XNLI). Other reasoning tasks may show different patterns.
• Consumer hardware: Training on Kaggle T4 (16GB) with 4-bit quantization introduces approximation that may affect adapter quality compared to full-precision training.

Related Resources

• Training data: legesher/language-decoded-data
• Community code: legesher/language-decoded-community
• Experiment tracking: legesher/language-decoded-experiments
• Transpilation tool: Legesher on GitHub

Citation

@misc{language-decoded-2026,
  title={Language Decoded: Investigating Language-Dependent vs. Structure-Dependent Reasoning Benefits of Code},
  author={Madison Edgar and Saad Ahmed Bazaz and Tom Sherborne and Rashik Shahjahan and Khojasteh Mirza and Sarah Jawaid and Rafay Mustafa and Sohaib Ahmed Bazaz},
  year={2026},
  publisher={Hugging Face},
  url={https://huggingface.co/legesher/language-decoded-lora}
}

License

Apache 2.0