Dataset Viewer
Auto-converted to Parquet Duplicate
question
stringlengths
138
142
metadata
stringlengths
108
116
task_source
stringclasses
1 value
formatted_prompt
listlengths
2
2
response
listlengths
4
4
response__extractions
listlengths
4
4
response__extractions_metadata
listlengths
4
4
eval_correct
listlengths
4
4
eval_correct__metadata
listlengths
4
4
Using ALL of the numbers 6, 85, 30, 99, create an expression that equals 441. You must use each number exactly once. You may use +, -, *, /.
{"numbers": [6, 85, 30, 99], "target": 441, "operators_used": ["*", "+", "-"], "expression": "6 * 85 + 30 - 99"}
countdown
[ { "content": "You are a math problem solver. Solve the given arithmetic problem step by step. Show your work and then provide your final answer inside <answer> tags.\n\nFor example: <answer>10 + 5 * 2 = 20</answer>\n\nThe answer should be a single arithmetic expression using the given numbers that equals the ta...
[ "To create an expression that equals 441 using all the numbers 6, 85, 30, and 99, let's analyze the possibilities:\n\n1. Expressions involving 6: Since 6 is the smallest number, it's likely that it will be multiplied by another number to get a larger result. \n2. Expressions involving 85: 85 is a large number, so i...
[ [], [], [], [] ]
[ [], [], [], [] ]
[ [], [], [], [] ]
[ [], [], [], [] ]
"Using ALL of the numbers 66, 64, 12, 1, create an expression that equals 142. You must use each num(...TRUNCATED)
"{\"numbers\": [66, 64, 12, 1], \"target\": 142, \"operators_used\": [\"+\", \"+\", \"/\"], \"expres(...TRUNCATED)
countdown
[{"content":"You are a math problem solver. Solve the given arithmetic problem step by step. Show yo(...TRUNCATED)
["To create an expression that equals 142 using each of the given numbers exactly once, let's try di(...TRUNCATED)
[ [], [], [ "130 * 1.1 = 143" ], [ "(64 + 66) * 1 + 12 = 142" ] ]
[[],[],[{"end_index":992,"end_phrase":"</answer>","start_index":960,"start_phrase":"<answer>"}],[{"e(...TRUNCATED)
[ [], [], [ false ], [ true ] ]
[[],[],[{"answer_block":"130 * 1.1 = 143","error":"Expression uses an unavailable or repeated number(...TRUNCATED)
"Using ALL of the numbers 9, 3, 49, 77, create an expression that equals 224. You must use each numb(...TRUNCATED)
"{\"numbers\": [9, 3, 49, 77], \"target\": 224, \"operators_used\": [\"/\", \"*\", \"+\"], \"express(...TRUNCATED)
countdown
[{"content":"You are a math problem solver. Solve the given arithmetic problem step by step. Show yo(...TRUNCATED)
["To create an expression that equals 224 using the given numbers, let's analyze the possibilities:\(...TRUNCATED)
[ [ "77 * 3 - 7" ], [], [], [] ]
[ [ { "end_index": 918, "end_phrase": "</answer>", "start_index": 891, "start_phrase": "<answer>" } ], [], [], [] ]
[ [ false ], [], [], [] ]
[[{"answer_block":"77 * 3 - 7","error":"Expression uses an unavailable or repeated number","final_an(...TRUNCATED)
"Using ALL of the numbers 99, 83, 11, 15, create an expression that equals 732. You must use each nu(...TRUNCATED)
"{\"numbers\": [99, 83, 11, 15], \"target\": 732, \"operators_used\": [\"*\", \"/\", \"-\"], \"expre(...TRUNCATED)
countdown
[{"content":"You are a math problem solver. Solve the given arithmetic problem step by step. Show yo(...TRUNCATED)
["To create an expression that equals 732 using all the given numbers, we can try different combinat(...TRUNCATED)
[ [], [], [], [] ]
[ [], [], [], [] ]
[ [], [], [], [] ]
[ [], [], [], [] ]
"Using ALL of the numbers 75, 62, 65, 68, create an expression that equals 146. You must use each nu(...TRUNCATED)
"{\"numbers\": [75, 62, 65, 68], \"target\": 146, \"operators_used\": [\"-\", \"+\", \"+\"], \"expre(...TRUNCATED)
countdown
[{"content":"You are a math problem solver. Solve the given arithmetic problem step by step. Show yo(...TRUNCATED)
["To create an expression that equals 146, we can try different combinations of the given numbers.\n(...TRUNCATED)
[ [], [], [], [] ]
[ [], [], [], [] ]
[ [], [], [], [] ]
[ [], [], [], [] ]
"Using ALL of the numbers 77, 1, 87, 31, create an expression that equals 133. You must use each num(...TRUNCATED)
"{\"numbers\": [77, 1, 87, 31], \"target\": 133, \"operators_used\": [\"/\", \"+\", \"-\"], \"expres(...TRUNCATED)
countdown
[{"content":"You are a math problem solver. Solve the given arithmetic problem step by step. Show yo(...TRUNCATED)
["Let's analyze the numbers and try to create an expression that equals 133. \n\nWe have 77, 1, 87, (...TRUNCATED)
[ [ "77 - 31 + 87 * 1 = 133" ], [], [ "46 + 87" ], [] ]
[[{"end_index":1577,"end_phrase":"</answer>","start_index":1538,"start_phrase":"<answer>"}],[],[{"en(...TRUNCATED)
[ [ true ], [], [ false ], [] ]
[[{"answer_block":"77 - 31 + 87 * 1 = 133","error":null,"final_answer":133.0,"is_correct":true}],[],(...TRUNCATED)
"Using ALL of the numbers 86, 92, 23, 31, create an expression that equals 313. You must use each nu(...TRUNCATED)
"{\"numbers\": [86, 92, 23, 31], \"target\": 313, \"operators_used\": [\"*\", \"/\", \"-\"], \"expre(...TRUNCATED)
countdown
[{"content":"You are a math problem solver. Solve the given arithmetic problem step by step. Show yo(...TRUNCATED)
["To solve this problem, we need to find an expression that equals 313 using the numbers 86, 92, 23,(...TRUNCATED)
[ [], [], [], [] ]
[ [], [], [], [] ]
[ [], [], [], [] ]
[ [], [], [], [] ]
"Using ALL of the numbers 65, 63, 1, 7, create an expression that equals 896. You must use each numb(...TRUNCATED)
"{\"numbers\": [65, 63, 1, 7], \"target\": 896, \"operators_used\": [\"+\", \"/\", \"*\"], \"express(...TRUNCATED)
countdown
[{"content":"You are a math problem solver. Solve the given arithmetic problem step by step. Show yo(...TRUNCATED)
["To create an expression that equals 896 using the given numbers 65, 63, 1, and 7, let's try differ(...TRUNCATED)
[ [ "7 * (63 - 1) + 7 * (1 * 64 + 1 * 32) = 896" ], [], [], [ "896 = (7 * 128) + 1" ] ]
[[{"end_index":3327,"end_phrase":"</answer>","start_index":3268,"start_phrase":"<answer>"}],[],[],[{(...TRUNCATED)
[ [ false ], [], [], [ false ] ]
[[{"answer_block":"7 * (63 - 1) + 7 * (1 * 64 + 1 * 32) = 896","error":"Expression does not equal st(...TRUNCATED)
"Using ALL of the numbers 27, 88, 32, 32, create an expression that equals 115. You must use each nu(...TRUNCATED)
"{\"numbers\": [27, 88, 32, 32], \"target\": 115, \"operators_used\": [\"+\", \"*\", \"/\"], \"expre(...TRUNCATED)
countdown
[{"content":"You are a math problem solver. Solve the given arithmetic problem step by step. Show yo(...TRUNCATED)
["To create an expression that equals 115 using the numbers 27, 88, 32, and 32, let's analyze the nu(...TRUNCATED)
[ [ "32 * (27/32) + 88 = 115" ], [], [ "32*32/3.75*3 + 27 = 115" ], [] ]
[[{"end_index":1423,"end_phrase":"</answer>","start_index":1383,"start_phrase":"<answer>"}],[],[{"en(...TRUNCATED)
[ [ true ], [], [ false ], [] ]
[[{"answer_block":"32 * (27/32) + 88 = 115","error":null,"final_answer":115.0,"is_correct":true}],[](...TRUNCATED)
"Using ALL of the numbers 95, 32, 35, 21, create an expression that equals 113. You must use each nu(...TRUNCATED)
"{\"numbers\": [95, 32, 35, 21], \"target\": 113, \"operators_used\": [\"+\", \"-\", \"+\"], \"expre(...TRUNCATED)
countdown
[{"content":"You are a math problem solver. Solve the given arithmetic problem step by step. Show yo(...TRUNCATED)
["To create an expression that equals 113 using each number exactly once, I'll try different combina(...TRUNCATED)
[ [], [], [], [] ]
[ [], [], [], [] ]
[ [], [], [], [] ]
[ [], [], [], [] ]
End of preview. Expand in Data Studio

t1-full-l3p2-3b-it-4arg-s1-base

Stage 1 checkpoint: base eval of hosted_vllm/meta-llama/Llama-3.2-3B-Instruct on 4-arg countdown (50 problems).

Dataset Info

  • Rows: 50
  • Columns: 9

Columns

Column Type Description
question Value('string') No description provided
metadata Value('string') No description provided
task_source Value('string') No description provided
formatted_prompt List({'content': Value('string'), 'role': Value('string')}) No description provided
response List(Value('string')) No description provided
response__extractions List(List(Value('string'))) No description provided
response__extractions_metadata List(List({'end_index': Value('int64'), 'end_phrase': Value('string'), 'start_index': Value('int64'), 'start_phrase': Value('string')})) No description provided
eval_correct List(List(Value('bool'))) No description provided
eval_correct__metadata List(List({'answer_block': Value('string'), 'error': Value('string'), 'final_answer': Value('float64'), 'is_correct': Value('bool')})) No description provided

Generation Parameters

{
  "script_name": "run_full_experiment.py",
  "model": "hosted_vllm/meta-llama/Llama-3.2-3B-Instruct",
  "description": "Stage 1 checkpoint: base eval of hosted_vllm/meta-llama/Llama-3.2-3B-Instruct on 4-arg countdown (50 problems).",
  "hyperparameters": {
    "temperature": 0.7,
    "top_p": 0.95,
    "max_tokens": 16384,
    "num_args": 4,
    "samples": 4
  },
  "custom_metadata": {
    "stage": "s1_base",
    "metrics": {
      "pass@1": 0.08,
      "pass@2": 0.12,
      "pass@3": 0.16,
      "pass@4": 0.18,
      "total_problems": 50,
      "samples_per_problem": 4
    }
  },
  "input_datasets": []
}

Usage

from datasets import load_dataset

dataset = load_dataset("reasoning-degeneration-dev/t1-full-l3p2-3b-it-4arg-s1-base", split="train")
print(f"Loaded {len(dataset)} rows")

This dataset is tracked in reasoning-degeneration-dev/PROJECT-MANIFEST

Downloads last month
5