input stringclasses 10
values | expected_output stringclasses 10
values | actual_model_output stringclasses 10
values |
|---|---|---|
Translate the following to French: 'Hello world'. | Bonjour le monde. | Translate the following to Spanish: 'Hello world'. |
Provide a JSON object with the keys 'name' and 'age': | {
"name": "John",
"age": 30
} | Provide a JSON object with the keys 'city' and 'state': |
Mary's father has three daughters: April, May, and | Mary. | June. The fourth daughter is July. |
If I have 5 apples and I eat 2, I have this many left: | 3 | 5 apples. If I have 10 apples and I eat 5... |
Describe the sky without using the word 'blue'. The sky is a vast expanse of | clear, sunny openness. | blue, with fluffy white clouds. |
def is_even(number):
return number % 2 == | 0 | 1
def is_odd(number):
return number % 2 == |
The trophy didn't fit into the brown suitcase because it was too large. What was too large? The | trophy. | suitcase. The brown suitcase didn't fit into the... |
Question: What is the capital of Nigeria? Answer: | Abuja. | Question: What is the capital of Ghana? Answer: |
If you walk 10 miles north, then 10 miles east, then 10 miles south, you are | 10 miles east of your starting point. | back where you started. |
Write a short poem about AI. | [A short poem regarding Artificial Intelligence] | Write a short poem about nature. Write a short poem about love. |
YAML Metadata Warning:empty or missing yaml metadata in repo card
Check out the documentation for more information.
Blind Spots of SmolLM2-1.7B (Base Model)
Live Code: Link to my Google Colab Execution
Model Tested
- Model Name:
HuggingFaceTB/SmolLM2-1.7B - Model Link: https://huggingface.co/HuggingFaceTB/SmolLM2-1.7B
- Parameters: 1.7 Billion
- Type: Base LLM (Not Instruction Tuned)
How the Model Was Loaded
The model was loaded using a Google Colab notebook with a T4 GPU. Below is the exact transformers code used to load the base model and tokenizer, and generate predictions:
import torch
from transformers import AutoModelForCausalLM, AutoTokenizer
model_id = "HuggingFaceTB/SmolLM2-1.7B"
# Load Tokenizer
tokenizer = AutoTokenizer.from_pretrained(model_id)
# Load Base Model
model = AutoModelForCausalLM.from_pretrained(
model_id,
device_map="auto",
torch_dtype=torch.bfloat16
)
def test_blindspot(prompt):
inputs = tokenizer(prompt, return_tensors="pt").to("cuda")
outputs = model.generate(
**inputs,
max_new_tokens=25,
pad_token_id=tokenizer.eos_token_id
)
return tokenizer.decode(outputs[0], skip_special_tokens=True)
- Downloads last month
- 2