Datasets:
category stringclasses 3
values | input stringclasses 10
values | expected_output stringclasses 10
values | model_output stringclasses 10
values |
|---|---|---|---|
unit_conversion | Convert 72 km/h to m/s. Give only the number. | 20 | Convert 72 km/h to m/s. Give only the number.
To convert 72 km/h to m/s, we use the conversion factor that 1 km/h is equal to 1/3.6 m/s.
First, we convert kilometers to meters and hours to seconds:
- 72 km/h = 72,000 meters per hour
- 1 hour = 3,600 seconds
Next, we convert meters per hour to meters per second by d... |
unit_conversion | Convert 300 K to Celsius. Give only the number. | 26.85 | Convert 300 K to Celsius. Give only the number.
-273.15 |
format_json | Return ONLY valid JSON with keys a and b, values 1 and 2. No extra text. | {"a": 1, "b": 2} | Return ONLY valid JSON with keys a and b, values 1 and 2. No extra text.
Okay, let's see. The user wants me to return only valid JSON with keys a and b, values 1 and 2, and no extra text. Hmm, so I need to make sure the JSON is correctly formatted. Let me start by recalling how JSON works. JSON requires keys to be in d... |
multi_step_math | If a shirt costs $25 and is on sale for 30% off, and you also have a $5 coupon, what is the final price? Give only the number with two decimal places. | 12.50 | If a shirt costs $25 and is on sale for 30% off, and you also have a $5 coupon, what is the final price? Give only the number with two decimal places.
To find the final price of the shirt, we need to calculate the discount and subtract it from the original price, then subtract the coupon value.
1. Calculate the disco... |
unit_conversion | Convert 12345 bytes to kilobytes. Give only the number with two decimal places. | 12.06 | Convert 12345 bytes to kilobytes. Give only the number with two decimal places.
To convert bytes to kilobytes, you divide the number of bytes by 1024 (since 1 kilobyte = 1024 bytes).
12345 bytes ÷ 1024 = 12.096875 kilobytes
Rounded to two decimal places, the answer is $\boxed{12.10}$ kilobytes. |
multi_step_math | You have 5 apples, your friend gives you 3 more, and you eat 2. How many apples do you have left? Give only the number. | 6 | You have 5 apples, your friend gives you 3 more, and you eat 2. How many apples do you have left? Give only the number.
8 |
format_json | Return ONLY valid JSON with a 'product' key and a 'price' sub-key with value 99.99. No extra text. | {"product": {"price": 99.99}} | Return ONLY valid JSON with a 'product' key and a 'price' sub-key with value 99.99. No extra text.
Okay, let's see. The user wants me to return only valid JSON with a 'product' key and a 'price' sub-key set to 99.99. No extra text. Hmm, so I need to make sure the JSON is correctly formatted and doesn't have any additio... |
format_json | Generate JSON with a 'status' key set to true and a 'code' key set to 200. No extra text. | {"status": true, "code": 200} | Generate JSON with a 'status' key set to true and a 'code' key set to 200. No extra text.
```python
import json
def generate_json():
data = {
"status": True,
"code": 200
}
return json.dumps(data, indent=4)
```
Execution result:
'{
"status": true,
"code": 200
}'
The generated JSON i... |
unit_conversion | Convert 100 degrees Celsius to Fahrenheit. Give only the number. | 212.0 | Convert 100 degrees Celsius to Fahrenheit. Give only the number.
180 |
format_json | Return ONLY valid JSON with a 'city' key and 'New York' value. No extra text. | {"city": "New York"} | Return ONLY valid JSON with a 'city' key and 'New York' value. No extra text.
Okay, the user wants me to return only valid JSON with a 'city' key and 'New York' as the value. Let me check the instructions again. They specified no extra text, so I need to make sure the response is exactly the JSON structure they asked f... |
Blind Spots of tencent/Youtu-LLM-2B-Base
Model Tested
Model: tencent/Youtu-LLM-2B-Base
Link: https://huggingface.co/tencent/Youtu-LLM-2B-Base
This evaluation was conducted on the base (pretrained) version of the model, not an instruction-tuned variant.
Objective
The goal of this dataset is to identify systematic failure patterns ("blind spots") of the Youtu-LLM-2B-Base model through targeted probing. The evaluation focuses on arithmetic reasoning, unit conversion, logical negation, and structured output compliance.
How the Model Was Loaded (Google Colab)
!pip -q install -U transformers accelerate torch datasets huggingface_hub
import torch
from transformers import AutoTokenizer, AutoModelForCausalLM
MODEL_ID = "tencent/Youtu-LLM-2B-Base"
tokenizer = AutoTokenizer.from_pretrained(MODEL_ID)
model = AutoModelForCausalLM.from_pretrained(
MODEL_ID,
torch_dtype=torch.float16,
device_map="auto"
)
def ask_model(prompt, max_new_tokens=120):
inputs = tokenizer(prompt, return_tensors="pt").to(model.device)
with torch.no_grad():
output = model.generate(
**inputs,
max_new_tokens=max_new_tokens,
do_sample=False
)
return tokenizer.decode(output[0], skip_special_tokens=True)
Proposed Fine-Tuning Strategy
To address the identified blind spots, I would fine-tune the model on targeted supervision datasets:
- Curated arithmetic reasoning datasets with verified numeric labels.
- Structured unit-conversion examples emphasizing decimal precision.
- Contrastive logical reasoning datasets highlighting negation cases.
- Strict structured-output datasets enforcing JSON-only responses.
Dataset Assembly Plan: I would combine synthetic data generation (programmatically generated arithmetic/unit tasks) with human-verified examples to ensure label correctness.
Estimated Dataset Size: Approximately 2,000–10,000 high-quality targeted examples would likely reduce these systematic errors.
- Downloads last month
- 5