datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
tiny-aya-l2-thinker-multilingual-reasoning
Tiny Aya L2 Multilingual Reasoning (44 languages)
Translated multilingual reasoning traces used to train Tiny Aya L2-Thinker.
Each example has the prompt, thinking, and answer in the same non-English language alongside the original texts in English.
Data source
Prompts from AM-DeepSeek-R1-0528-Distilled
Thinking traces and outputs distilled from gpt-oss-120b
Translated with command-a-translate and DeepSeek-V3
Languages (44)
Language
Train… See the full description on the dataset page: https://huggingface.co/datasets/CohereLabs/tiny-aya-l2-thinker-multilingual-reasoning.tiny-aya-l2-thinker-multilingual-reasoning
Tiny Aya L2 Multilingual Reasoning (44 languages)
Translated multilingual reasoning traces used to train Tiny Aya L2-Thinker.
Each example has the prompt, thinking, and answer in the same non-English language alongside the original texts in English.
Languages (44)
Language
Train
Test
Total
Amharic (am)
3,807
448
4,255
Arabic (ar)
22,968
2,538
25,506
Bulgarian (bg)
4,177
452
4,629
Bengali (bn)
3,803
422
4,225
Catalan (ca)
4,251
512
4,763
Czech… See the full description on the dataset page: https://huggingface.co/datasets/erenyeager-1/tiny-aya-l2-thinker-multilingual-reasoning.sorry-bench-202503-multilingual
sorry-bench-202503-multilingual
Multilingual version of SorryBench — a benchmark for evaluating LLM safety refusals across 44 harm categories and 21 prompt styles.
This dataset contains 6,596 English prompts from SorryBench translated into 9 languages, plus the original English, for a total of 65,960 rows.
Schema
Column
Type
Description
question_id
int
Original SorryBench question ID
category
int
Harm category (1-44)
prompt_style
string
SorryBench prompt… See the full description on the dataset page: https://huggingface.co/datasets/tiny-aya-safety/sorry-bench-202503-multilingual.fusion-aya-math-bench
Dataset Card for Fusion Aya Math Bench
Summary
Fusion Aya Math Bench is a multilingual, olympiad-level mathematical reasoning dataset. Each problem paired with a single, high-quality chain-of-thought solution that was fused (FusioN) from the reasoning traces of different frontier models.
Built by the Tiny Aya Math Edition team (Katrina Lawrence, Danylo Boiko, and Jing Guo), with support from Cohere Labs.
Pipeline
Derived from the open-ended… See the full description on the dataset page: https://huggingface.co/datasets/tiny-aya-math-edition/fusion-aya-math-bench.tiny-aya-global-blindspots
tiny-aya-global: Blind Spot Evaluation
Model: CohereLabs/tiny-aya-global — ~3B parameter multilingual conversational model
Dataset: kelvinyelyen/tiny-aya-global-blindspots
Ten targeted probes designed to surface specific failure mechanisms, not aggregate accuracy. Each probe was run once under greedy decoding (temperature=0.0), then re-run 5x under sampling (temperature=0.7) to check whether each failure is a stable pattern or a one-off. Result: 7 clear failures, 1 pass, 1… See the full description on the dataset page: https://huggingface.co/datasets/kelvinyelyen/tiny-aya-global-blindspots.tiny-aya-base-blind-spots
Tiny Aya Base — Blind Spots Dataset
Overview
This dataset documents blind spots identified in CohereLabs/tiny-aya-base, a multilingual base language model (3.35B parameters, 70+ languages). Each entry contains a prompt, the expected correct output, the model's actual output, and a human annotation of the error type.
The model scored 5/18 (28%) on our evaluation prompts.
Categories Tested
Multilingual (6 prompts, 2 correct): Yoruba, Igbo, Hausa translation… See the full description on the dataset page: https://huggingface.co/datasets/Ifihan/tiny-aya-base-blind-spots.tiny-aya-base-blindspots
Blind Spots: CohereLabs/tiny-aya-base
Model Tested
CohereLabs/tiny-aya-base
Property
Value
Parameters
3.35 billion (BF16)
Architecture
Cohere2ForCausalLM
Type
Pure pre-trained base model (not SFT/RLHF)
Languages
70+ languages
Released
February 13, 2026
License
CC-BY-NC-4.0
Context
8K input / 8K output
Access
Gated (agree to share contact info)
Why this model?
Tiny Aya is Cohere Labs' open-weights pre-trained 3.35B parameter base… See the full description on the dataset page: https://huggingface.co/datasets/mozayed/tiny-aya-base-blindspots.tiny-aya-base-blind-spots
tiny-aya-base Blind Spot Dataset
Model Tested
CohereLabs/tiny-aya-base
Architecture: Cohere2
Parameters: 3B
Type: Base model (not instruction-tuned)
Languages: 70+ languages
Released: March 2026
How I Loaded the Model
!pip install transformers torch accelerate -q
from transformers import AutoModelForCausalLM, AutoTokenizer
from huggingface_hub import login
import torch
login() # HuggingFace token required (model is gated)
model_name =… See the full description on the dataset page: https://huggingface.co/datasets/JackRabbit1122/tiny-aya-base-blind-spots.tiny-aya-translate-hinglish-casual-stripped
Dataset Card for tiny-aya-translate-hinglish-casual-stripped
Dataset Summary
tiny-aya-translate-hinglish-casual-stripped is a lightweight, text-only derivative of the original tiny-aya-translate/hinglish-casual dataset.
The original dataset is designed for simultaneous translation and contains many columns including audio references, speaker metadata, and duration. It also includes paralinguistic tags (e.g., <sigh>, <laugh>, <chuckle>) embedded within the… See the full description on the dataset page: https://huggingface.co/datasets/bingbangboom/tiny-aya-translate-hinglish-casual-stripped.tiny-aya-base-blind-spots
Blind Spots of a Frontier Base Model: Evaluation Dataset
This dataset documents blind spots discovered in a frontier open-weight base model through 19 structured evaluation tests. It was assembled as part of an assignment on identifying model weaknesses using the HelloBench evaluation framework.
Model Tested
CohereLabs/tiny-aya-base
Architecture: Transformer with Sliding Window Attention (SWA) (window size 4096, with RoPE) on three layers + one global attention layer… See the full description on the dataset page: https://huggingface.co/datasets/Mawube/tiny-aya-base-blind-spots.tiny-aya-global-evaluation
Tiny-Aya-Global Reasoning Blind Spots (TAG-RBS)
This diagnostic dataset identifies the logical, mathematical, and constraint-satisfaction "blind spots" of the Tiny-Aya-Global (3.35B) model. It was manually constructed to test the boundary conditions of compact multilingual models and evaluate their susceptibility to post-hoc rationalization.
Dataset Overview
Dataset Size: 50 hand-crafted prompts.
Evaluation Target: CohereLabs/tiny-aya-global (3.35B parameters).… See the full description on the dataset page: https://huggingface.co/datasets/yonasachule/tiny-aya-global-evaluation.tiny-aya-base-blind-spots
tiny-aya-base — Blind Spots Dataset
Model Tested
CohereLabs/tiny-aya-base
Architecture: Cohere2 (Command-R style)
Parameters: ~3B
Released: February 2025
Type: Base model (not instruction-tuned). The chat variants tiny-aya-global,tiny-aya-earth, tiny-aya-water, and tiny-aya-fire are fine-tuned from this base.
How the Model Was Loaded
Tested on Google Colab (free T4 GPU, ~15 GB VRAM).
import torch
from transformers import AutoModelForCausalLM… See the full description on the dataset page: https://huggingface.co/datasets/mohamedstohy/tiny-aya-base-blind-spots.tiny-aya-blind-spots
Dataset: Tiny-Aya-Base Blind Spots
This dataset was created as part of a technical challenge to identify the blind spots of the models. It specifically targets CohereLabs/tiny-aya-base. The model is a 3.35B parameter multilingual base model released in early 2026.
Model Tested
Model: CohereLabs/tiny-aya-base
Parameters: 3.35 Billion
Modality: Text
How the Model was Loaded
The model was loaded using the transformers library on a Google Colab T4 GPU.… See the full description on the dataset page: https://huggingface.co/datasets/osamaahmed17/tiny-aya-blind-spots.ignatius-tiny-aya-analysis
Tiny Aya Base — Failure Analysis Dataset
Model Tested
Model: CohereLabs/tiny-aya-base
Parameters: 3.35B
Released: February 17, 2026
Architecture: Dense decoder-only Transformer, pretrained on 6T tokens across 70+ languages
How I Loaded the Model
I used Modal (free tier, T4 GPU — 16GB VRAM) to run inference.
import modal
app = modal.App("tiny-aya-probe")
image = modal.Image.debian_slim(python_version="3.11").pip_install(
"torch", "transformers>=4.51.0"… See the full description on the dataset page: https://huggingface.co/datasets/IgnatiusBalayo2024/ignatius-tiny-aya-analysis.tiny-aya-base-blindspots
Dataset Card: tiny-aya-base-blindspots
Summary
This is a dataset identifies structural, logical, and alignment blind spots in the CohereLabs/tiny-aya-base foundation model. It was created with fun, I have tried over a hundred examples and find out the model is solid for its size,
however, because it is a base model, It has hard time following instrucitons, for example even when i tell the model explicity "this is a fill the gap question", it introprates it as a MCQ… See the full description on the dataset page: https://huggingface.co/datasets/mansouro/tiny-aya-base-blindspots.