data_source
stringclasses
2 values
prompt
listlengths
2
2
ability
stringclasses
1 value
reward_model
dict
extra_info
dict
metadata
dict
nq
[ { "content": "You are a helpful and harmless assistant.", "role": "system" }, { "content": "Answer the given question step by step.\n Instructions:\n 1. First, conduct reasoning inside <think> and </think> tags whenever you receive new information.\n 2. If you need external knowledge, you can search using <search>query</search>.\n 3. When you receive search results, evaluate their usefulness and put your judgment inside <judge>Yes</judge> or <judge>No</judge> tags.\n 4. Based on your judgment, follow these strict rules:\n a. If the information is useful AND you now have sufficient information to provide a complete final answer, proceed directly to step 5.\n b. If the information is useful BUT you still need more details, you MUST search again with <search>...</search>.\n c. If the information is not useful, you MUST search again with <search>...</search>. You MUST NOT provide an answer in this case.\n 5. Provide your final answer in <answer>...</answer> tags. The <answer> tag marks the end of the task. After providing the <answer>, you MUST stop and generate no further text.\n\n Question: how many calories does a cosmic brownie have?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "280" ] }, "style": "rule" }
{ "index": 400, "need_tools_kwargs": true, "question": "how many calories does a cosmic brownie have?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "280" ] }, "question": "how many calories does a cosmic brownie have?" } } } }
null
nq
[ { "content": "You are a helpful and harmless assistant.", "role": "system" }, { "content": "Answer the given question step by step.\n Instructions:\n 1. First, conduct reasoning inside <think> and </think> tags whenever you receive new information.\n 2. If you need external knowledge, you can search using <search>query</search>.\n 3. When you receive search results, evaluate their usefulness and put your judgment inside <judge>Yes</judge> or <judge>No</judge> tags.\n 4. Based on your judgment, follow these strict rules:\n a. If the information is useful AND you now have sufficient information to provide a complete final answer, proceed directly to step 5.\n b. If the information is useful BUT you still need more details, you MUST search again with <search>...</search>.\n c. If the information is not useful, you MUST search again with <search>...</search>. You MUST NOT provide an answer in this case.\n 5. Provide your final answer in <answer>...</answer> tags. The <answer> tag marks the end of the task. After providing the <answer>, you MUST stop and generate no further text.\n\n Question: how many episodes in season 3 of red oaks?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "6" ] }, "style": "rule" }
{ "index": 401, "need_tools_kwargs": true, "question": "how many episodes in season 3 of red oaks?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "6" ] }, "question": "how many episodes in season 3 of red oaks?" } } } }
null
nq
[ { "content": "You are a helpful and harmless assistant.", "role": "system" }, { "content": "Answer the given question step by step.\n Instructions:\n 1. First, conduct reasoning inside <think> and </think> tags whenever you receive new information.\n 2. If you need external knowledge, you can search using <search>query</search>.\n 3. When you receive search results, evaluate their usefulness and put your judgment inside <judge>Yes</judge> or <judge>No</judge> tags.\n 4. Based on your judgment, follow these strict rules:\n a. If the information is useful AND you now have sufficient information to provide a complete final answer, proceed directly to step 5.\n b. If the information is useful BUT you still need more details, you MUST search again with <search>...</search>.\n c. If the information is not useful, you MUST search again with <search>...</search>. You MUST NOT provide an answer in this case.\n 5. Provide your final answer in <answer>...</answer> tags. The <answer> tag marks the end of the task. After providing the <answer>, you MUST stop and generate no further text.\n\n Question: when did dustin johnson win the us open?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "2016" ] }, "style": "rule" }
{ "index": 402, "need_tools_kwargs": true, "question": "when did dustin johnson win the us open?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "2016" ] }, "question": "when did dustin johnson win the us open?" } } } }
null
nq
[ { "content": "You are a helpful and harmless assistant.", "role": "system" }, { "content": "Answer the given question step by step.\n Instructions:\n 1. First, conduct reasoning inside <think> and </think> tags whenever you receive new information.\n 2. If you need external knowledge, you can search using <search>query</search>.\n 3. When you receive search results, evaluate their usefulness and put your judgment inside <judge>Yes</judge> or <judge>No</judge> tags.\n 4. Based on your judgment, follow these strict rules:\n a. If the information is useful AND you now have sufficient information to provide a complete final answer, proceed directly to step 5.\n b. If the information is useful BUT you still need more details, you MUST search again with <search>...</search>.\n c. If the information is not useful, you MUST search again with <search>...</search>. You MUST NOT provide an answer in this case.\n 5. Provide your final answer in <answer>...</answer> tags. The <answer> tag marks the end of the task. After providing the <answer>, you MUST stop and generate no further text.\n\n Question: when will new 20 dollar bill be released?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "by 2020" ] }, "style": "rule" }
{ "index": 403, "need_tools_kwargs": true, "question": "when will new 20 dollar bill be released?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "by 2020" ] }, "question": "when will new 20 dollar bill be released?" } } } }
null
nq
[ { "content": "You are a helpful and harmless assistant.", "role": "system" }, { "content": "Answer the given question step by step.\n Instructions:\n 1. First, conduct reasoning inside <think> and </think> tags whenever you receive new information.\n 2. If you need external knowledge, you can search using <search>query</search>.\n 3. When you receive search results, evaluate their usefulness and put your judgment inside <judge>Yes</judge> or <judge>No</judge> tags.\n 4. Based on your judgment, follow these strict rules:\n a. If the information is useful AND you now have sufficient information to provide a complete final answer, proceed directly to step 5.\n b. If the information is useful BUT you still need more details, you MUST search again with <search>...</search>.\n c. If the information is not useful, you MUST search again with <search>...</search>. You MUST NOT provide an answer in this case.\n 5. Provide your final answer in <answer>...</answer> tags. The <answer> tag marks the end of the task. After providing the <answer>, you MUST stop and generate no further text.\n\n Question: who was the last republican mayor in chicago?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "William Hale Thompson" ] }, "style": "rule" }
{ "index": 404, "need_tools_kwargs": true, "question": "who was the last republican mayor in chicago?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "William Hale Thompson" ] }, "question": "who was the last republican mayor in chicago?" } } } }
null
nq
[ { "content": "You are a helpful and harmless assistant.", "role": "system" }, { "content": "Answer the given question step by step.\n Instructions:\n 1. First, conduct reasoning inside <think> and </think> tags whenever you receive new information.\n 2. If you need external knowledge, you can search using <search>query</search>.\n 3. When you receive search results, evaluate their usefulness and put your judgment inside <judge>Yes</judge> or <judge>No</judge> tags.\n 4. Based on your judgment, follow these strict rules:\n a. If the information is useful AND you now have sufficient information to provide a complete final answer, proceed directly to step 5.\n b. If the information is useful BUT you still need more details, you MUST search again with <search>...</search>.\n c. If the information is not useful, you MUST search again with <search>...</search>. You MUST NOT provide an answer in this case.\n 5. Provide your final answer in <answer>...</answer> tags. The <answer> tag marks the end of the task. After providing the <answer>, you MUST stop and generate no further text.\n\n Question: who played the voice of jabba the hutt?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "Larry Ward" ] }, "style": "rule" }
{ "index": 405, "need_tools_kwargs": true, "question": "who played the voice of jabba the hutt?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "Larry Ward" ] }, "question": "who played the voice of jabba the hutt?" } } } }
null
nq
[ { "content": "You are a helpful and harmless assistant.", "role": "system" }, { "content": "Answer the given question step by step.\n Instructions:\n 1. First, conduct reasoning inside <think> and </think> tags whenever you receive new information.\n 2. If you need external knowledge, you can search using <search>query</search>.\n 3. When you receive search results, evaluate their usefulness and put your judgment inside <judge>Yes</judge> or <judge>No</judge> tags.\n 4. Based on your judgment, follow these strict rules:\n a. If the information is useful AND you now have sufficient information to provide a complete final answer, proceed directly to step 5.\n b. If the information is useful BUT you still need more details, you MUST search again with <search>...</search>.\n c. If the information is not useful, you MUST search again with <search>...</search>. You MUST NOT provide an answer in this case.\n 5. Provide your final answer in <answer>...</answer> tags. The <answer> tag marks the end of the task. After providing the <answer>, you MUST stop and generate no further text.\n\n Question: who won men's figure skating in olympics 2018?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "Yuzuru Hanyu" ] }, "style": "rule" }
{ "index": 406, "need_tools_kwargs": true, "question": "who won men's figure skating in olympics 2018?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "Yuzuru Hanyu" ] }, "question": "who won men's figure skating in olympics 2018?" } } } }
null
nq
[ { "content": "You are a helpful and harmless assistant.", "role": "system" }, { "content": "Answer the given question step by step.\n Instructions:\n 1. First, conduct reasoning inside <think> and </think> tags whenever you receive new information.\n 2. If you need external knowledge, you can search using <search>query</search>.\n 3. When you receive search results, evaluate their usefulness and put your judgment inside <judge>Yes</judge> or <judge>No</judge> tags.\n 4. Based on your judgment, follow these strict rules:\n a. If the information is useful AND you now have sufficient information to provide a complete final answer, proceed directly to step 5.\n b. If the information is useful BUT you still need more details, you MUST search again with <search>...</search>.\n c. If the information is not useful, you MUST search again with <search>...</search>. You MUST NOT provide an answer in this case.\n 5. Provide your final answer in <answer>...</answer> tags. The <answer> tag marks the end of the task. After providing the <answer>, you MUST stop and generate no further text.\n\n Question: who plays dennis on it's always sunny?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "Glenn Franklin Howerton III" ] }, "style": "rule" }
{ "index": 407, "need_tools_kwargs": true, "question": "who plays dennis on it's always sunny?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "Glenn Franklin Howerton III" ] }, "question": "who plays dennis on it's always sunny?" } } } }
null
nq
[ { "content": "You are a helpful and harmless assistant.", "role": "system" }, { "content": "Answer the given question step by step.\n Instructions:\n 1. First, conduct reasoning inside <think> and </think> tags whenever you receive new information.\n 2. If you need external knowledge, you can search using <search>query</search>.\n 3. When you receive search results, evaluate their usefulness and put your judgment inside <judge>Yes</judge> or <judge>No</judge> tags.\n 4. Based on your judgment, follow these strict rules:\n a. If the information is useful AND you now have sufficient information to provide a complete final answer, proceed directly to step 5.\n b. If the information is useful BUT you still need more details, you MUST search again with <search>...</search>.\n c. If the information is not useful, you MUST search again with <search>...</search>. You MUST NOT provide an answer in this case.\n 5. Provide your final answer in <answer>...</answer> tags. The <answer> tag marks the end of the task. After providing the <answer>, you MUST stop and generate no further text.\n\n Question: when did the battle of omaha beach start?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "June 6, 1944" ] }, "style": "rule" }
{ "index": 408, "need_tools_kwargs": true, "question": "when did the battle of omaha beach start?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "June 6, 1944" ] }, "question": "when did the battle of omaha beach start?" } } } }
null
nq
[ { "content": "You are a helpful and harmless assistant.", "role": "system" }, { "content": "Answer the given question step by step.\n Instructions:\n 1. First, conduct reasoning inside <think> and </think> tags whenever you receive new information.\n 2. If you need external knowledge, you can search using <search>query</search>.\n 3. When you receive search results, evaluate their usefulness and put your judgment inside <judge>Yes</judge> or <judge>No</judge> tags.\n 4. Based on your judgment, follow these strict rules:\n a. If the information is useful AND you now have sufficient information to provide a complete final answer, proceed directly to step 5.\n b. If the information is useful BUT you still need more details, you MUST search again with <search>...</search>.\n c. If the information is not useful, you MUST search again with <search>...</search>. You MUST NOT provide an answer in this case.\n 5. Provide your final answer in <answer>...</answer> tags. The <answer> tag marks the end of the task. After providing the <answer>, you MUST stop and generate no further text.\n\n Question: which nfl team is responsible for the 1985 shuffle?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "Chicago Bears" ] }, "style": "rule" }
{ "index": 409, "need_tools_kwargs": true, "question": "which nfl team is responsible for the 1985 shuffle?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "Chicago Bears" ] }, "question": "which nfl team is responsible for the 1985 shuffle?" } } } }
null
nq
[ { "content": "You are a helpful and harmless assistant.", "role": "system" }, { "content": "Answer the given question step by step.\n Instructions:\n 1. First, conduct reasoning inside <think> and </think> tags whenever you receive new information.\n 2. If you need external knowledge, you can search using <search>query</search>.\n 3. When you receive search results, evaluate their usefulness and put your judgment inside <judge>Yes</judge> or <judge>No</judge> tags.\n 4. Based on your judgment, follow these strict rules:\n a. If the information is useful AND you now have sufficient information to provide a complete final answer, proceed directly to step 5.\n b. If the information is useful BUT you still need more details, you MUST search again with <search>...</search>.\n c. If the information is not useful, you MUST search again with <search>...</search>. You MUST NOT provide an answer in this case.\n 5. Provide your final answer in <answer>...</answer> tags. The <answer> tag marks the end of the task. After providing the <answer>, you MUST stop and generate no further text.\n\n Question: who won the women's fifa world cup?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "United States" ] }, "style": "rule" }
{ "index": 410, "need_tools_kwargs": true, "question": "who won the women's fifa world cup?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "United States" ] }, "question": "who won the women's fifa world cup?" } } } }
null
nq
[ { "content": "You are a helpful and harmless assistant.", "role": "system" }, { "content": "Answer the given question step by step.\n Instructions:\n 1. First, conduct reasoning inside <think> and </think> tags whenever you receive new information.\n 2. If you need external knowledge, you can search using <search>query</search>.\n 3. When you receive search results, evaluate their usefulness and put your judgment inside <judge>Yes</judge> or <judge>No</judge> tags.\n 4. Based on your judgment, follow these strict rules:\n a. If the information is useful AND you now have sufficient information to provide a complete final answer, proceed directly to step 5.\n b. If the information is useful BUT you still need more details, you MUST search again with <search>...</search>.\n c. If the information is not useful, you MUST search again with <search>...</search>. You MUST NOT provide an answer in this case.\n 5. Provide your final answer in <answer>...</answer> tags. The <answer> tag marks the end of the task. After providing the <answer>, you MUST stop and generate no further text.\n\n Question: last team to win fa cup not in top division?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "West Ham United" ] }, "style": "rule" }
{ "index": 411, "need_tools_kwargs": true, "question": "last team to win fa cup not in top division?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "West Ham United" ] }, "question": "last team to win fa cup not in top division?" } } } }
null
nq
[ { "content": "You are a helpful and harmless assistant.", "role": "system" }, { "content": "Answer the given question step by step.\n Instructions:\n 1. First, conduct reasoning inside <think> and </think> tags whenever you receive new information.\n 2. If you need external knowledge, you can search using <search>query</search>.\n 3. When you receive search results, evaluate their usefulness and put your judgment inside <judge>Yes</judge> or <judge>No</judge> tags.\n 4. Based on your judgment, follow these strict rules:\n a. If the information is useful AND you now have sufficient information to provide a complete final answer, proceed directly to step 5.\n b. If the information is useful BUT you still need more details, you MUST search again with <search>...</search>.\n c. If the information is not useful, you MUST search again with <search>...</search>. You MUST NOT provide an answer in this case.\n 5. Provide your final answer in <answer>...</answer> tags. The <answer> tag marks the end of the task. After providing the <answer>, you MUST stop and generate no further text.\n\n Question: what is the rank of india in the richest country?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "world, ranked 130", "Asia: ranked 2" ] }, "style": "rule" }
{ "index": 412, "need_tools_kwargs": true, "question": "what is the rank of india in the richest country?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "world, ranked 130", "Asia: ranked 2" ] }, "question": "what is the rank of india in the richest country?" } } } }
null
nq
[ { "content": "You are a helpful and harmless assistant.", "role": "system" }, { "content": "Answer the given question step by step.\n Instructions:\n 1. First, conduct reasoning inside <think> and </think> tags whenever you receive new information.\n 2. If you need external knowledge, you can search using <search>query</search>.\n 3. When you receive search results, evaluate their usefulness and put your judgment inside <judge>Yes</judge> or <judge>No</judge> tags.\n 4. Based on your judgment, follow these strict rules:\n a. If the information is useful AND you now have sufficient information to provide a complete final answer, proceed directly to step 5.\n b. If the information is useful BUT you still need more details, you MUST search again with <search>...</search>.\n c. If the information is not useful, you MUST search again with <search>...</search>. You MUST NOT provide an answer in this case.\n 5. Provide your final answer in <answer>...</answer> tags. The <answer> tag marks the end of the task. After providing the <answer>, you MUST stop and generate no further text.\n\n Question: what is the best selling chocolate in the world?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "Reese's Peanut Butter Cups" ] }, "style": "rule" }
{ "index": 413, "need_tools_kwargs": true, "question": "what is the best selling chocolate in the world?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "Reese's Peanut Butter Cups" ] }, "question": "what is the best selling chocolate in the world?" } } } }
null
nq
[ { "content": "You are a helpful and harmless assistant.", "role": "system" }, { "content": "Answer the given question step by step.\n Instructions:\n 1. First, conduct reasoning inside <think> and </think> tags whenever you receive new information.\n 2. If you need external knowledge, you can search using <search>query</search>.\n 3. When you receive search results, evaluate their usefulness and put your judgment inside <judge>Yes</judge> or <judge>No</judge> tags.\n 4. Based on your judgment, follow these strict rules:\n a. If the information is useful AND you now have sufficient information to provide a complete final answer, proceed directly to step 5.\n b. If the information is useful BUT you still need more details, you MUST search again with <search>...</search>.\n c. If the information is not useful, you MUST search again with <search>...</search>. You MUST NOT provide an answer in this case.\n 5. Provide your final answer in <answer>...</answer> tags. The <answer> tag marks the end of the task. After providing the <answer>, you MUST stop and generate no further text.\n\n Question: who was the 41st president of the united states of america?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "George Herbert Walker Bush" ] }, "style": "rule" }
{ "index": 414, "need_tools_kwargs": true, "question": "who was the 41st president of the united states of america?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "George Herbert Walker Bush" ] }, "question": "who was the 41st president of the united states of america?" } } } }
null
nq
[ { "content": "You are a helpful and harmless assistant.", "role": "system" }, { "content": "Answer the given question step by step.\n Instructions:\n 1. First, conduct reasoning inside <think> and </think> tags whenever you receive new information.\n 2. If you need external knowledge, you can search using <search>query</search>.\n 3. When you receive search results, evaluate their usefulness and put your judgment inside <judge>Yes</judge> or <judge>No</judge> tags.\n 4. Based on your judgment, follow these strict rules:\n a. If the information is useful AND you now have sufficient information to provide a complete final answer, proceed directly to step 5.\n b. If the information is useful BUT you still need more details, you MUST search again with <search>...</search>.\n c. If the information is not useful, you MUST search again with <search>...</search>. You MUST NOT provide an answer in this case.\n 5. Provide your final answer in <answer>...</answer> tags. The <answer> tag marks the end of the task. After providing the <answer>, you MUST stop and generate no further text.\n\n Question: an alpha particle has a mass number of?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "4u" ] }, "style": "rule" }
{ "index": 415, "need_tools_kwargs": true, "question": "an alpha particle has a mass number of?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "4u" ] }, "question": "an alpha particle has a mass number of?" } } } }
null
nq
[ { "content": "You are a helpful and harmless assistant.", "role": "system" }, { "content": "Answer the given question step by step.\n Instructions:\n 1. First, conduct reasoning inside <think> and </think> tags whenever you receive new information.\n 2. If you need external knowledge, you can search using <search>query</search>.\n 3. When you receive search results, evaluate their usefulness and put your judgment inside <judge>Yes</judge> or <judge>No</judge> tags.\n 4. Based on your judgment, follow these strict rules:\n a. If the information is useful AND you now have sufficient information to provide a complete final answer, proceed directly to step 5.\n b. If the information is useful BUT you still need more details, you MUST search again with <search>...</search>.\n c. If the information is not useful, you MUST search again with <search>...</search>. You MUST NOT provide an answer in this case.\n 5. Provide your final answer in <answer>...</answer> tags. The <answer> tag marks the end of the task. After providing the <answer>, you MUST stop and generate no further text.\n\n Question: who plays ruben amaro jr on the goldbergs?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "Niko Guardado" ] }, "style": "rule" }
{ "index": 416, "need_tools_kwargs": true, "question": "who plays ruben amaro jr on the goldbergs?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "Niko Guardado" ] }, "question": "who plays ruben amaro jr on the goldbergs?" } } } }
null
nq
[ { "content": "You are a helpful and harmless assistant.", "role": "system" }, { "content": "Answer the given question step by step.\n Instructions:\n 1. First, conduct reasoning inside <think> and </think> tags whenever you receive new information.\n 2. If you need external knowledge, you can search using <search>query</search>.\n 3. When you receive search results, evaluate their usefulness and put your judgment inside <judge>Yes</judge> or <judge>No</judge> tags.\n 4. Based on your judgment, follow these strict rules:\n a. If the information is useful AND you now have sufficient information to provide a complete final answer, proceed directly to step 5.\n b. If the information is useful BUT you still need more details, you MUST search again with <search>...</search>.\n c. If the information is not useful, you MUST search again with <search>...</search>. You MUST NOT provide an answer in this case.\n 5. Provide your final answer in <answer>...</answer> tags. The <answer> tag marks the end of the task. After providing the <answer>, you MUST stop and generate no further text.\n\n Question: when is the last episode of grey's anatomy season 14?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "May 17, 2018" ] }, "style": "rule" }
{ "index": 417, "need_tools_kwargs": true, "question": "when is the last episode of grey's anatomy season 14?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "May 17, 2018" ] }, "question": "when is the last episode of grey's anatomy season 14?" } } } }
null
nq
[ { "content": "You are a helpful and harmless assistant.", "role": "system" }, { "content": "Answer the given question step by step.\n Instructions:\n 1. First, conduct reasoning inside <think> and </think> tags whenever you receive new information.\n 2. If you need external knowledge, you can search using <search>query</search>.\n 3. When you receive search results, evaluate their usefulness and put your judgment inside <judge>Yes</judge> or <judge>No</judge> tags.\n 4. Based on your judgment, follow these strict rules:\n a. If the information is useful AND you now have sufficient information to provide a complete final answer, proceed directly to step 5.\n b. If the information is useful BUT you still need more details, you MUST search again with <search>...</search>.\n c. If the information is not useful, you MUST search again with <search>...</search>. You MUST NOT provide an answer in this case.\n 5. Provide your final answer in <answer>...</answer> tags. The <answer> tag marks the end of the task. After providing the <answer>, you MUST stop and generate no further text.\n\n Question: what are the three branches of us government?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "legislative", "executive", "judicial" ] }, "style": "rule" }
{ "index": 418, "need_tools_kwargs": true, "question": "what are the three branches of us government?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "legislative", "executive", "judicial" ] }, "question": "what are the three branches of us government?" } } } }
null
nq
[ { "content": "You are a helpful and harmless assistant.", "role": "system" }, { "content": "Answer the given question step by step.\n Instructions:\n 1. First, conduct reasoning inside <think> and </think> tags whenever you receive new information.\n 2. If you need external knowledge, you can search using <search>query</search>.\n 3. When you receive search results, evaluate their usefulness and put your judgment inside <judge>Yes</judge> or <judge>No</judge> tags.\n 4. Based on your judgment, follow these strict rules:\n a. If the information is useful AND you now have sufficient information to provide a complete final answer, proceed directly to step 5.\n b. If the information is useful BUT you still need more details, you MUST search again with <search>...</search>.\n c. If the information is not useful, you MUST search again with <search>...</search>. You MUST NOT provide an answer in this case.\n 5. Provide your final answer in <answer>...</answer> tags. The <answer> tag marks the end of the task. After providing the <answer>, you MUST stop and generate no further text.\n\n Question: who won first womens money in the bank?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "Carmella" ] }, "style": "rule" }
{ "index": 419, "need_tools_kwargs": true, "question": "who won first womens money in the bank?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "Carmella" ] }, "question": "who won first womens money in the bank?" } } } }
null
nq
[ { "content": "You are a helpful and harmless assistant.", "role": "system" }, { "content": "Answer the given question step by step.\n Instructions:\n 1. First, conduct reasoning inside <think> and </think> tags whenever you receive new information.\n 2. If you need external knowledge, you can search using <search>query</search>.\n 3. When you receive search results, evaluate their usefulness and put your judgment inside <judge>Yes</judge> or <judge>No</judge> tags.\n 4. Based on your judgment, follow these strict rules:\n a. If the information is useful AND you now have sufficient information to provide a complete final answer, proceed directly to step 5.\n b. If the information is useful BUT you still need more details, you MUST search again with <search>...</search>.\n c. If the information is not useful, you MUST search again with <search>...</search>. You MUST NOT provide an answer in this case.\n 5. Provide your final answer in <answer>...</answer> tags. The <answer> tag marks the end of the task. After providing the <answer>, you MUST stop and generate no further text.\n\n Question: fear the walking dead season 3 new characters?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "Lisandra Tena", "Dayton Callie", "Daniel Sharman", "Sam Underwood" ] }, "style": "rule" }
{ "index": 420, "need_tools_kwargs": true, "question": "fear the walking dead season 3 new characters?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "Lisandra Tena", "Dayton Callie", "Daniel Sharman", "Sam Underwood" ] }, "question": "fear the walking dead season 3 new characters?" } } } }
null
nq
[ { "content": "You are a helpful and harmless assistant.", "role": "system" }, { "content": "Answer the given question step by step.\n Instructions:\n 1. First, conduct reasoning inside <think> and </think> tags whenever you receive new information.\n 2. If you need external knowledge, you can search using <search>query</search>.\n 3. When you receive search results, evaluate their usefulness and put your judgment inside <judge>Yes</judge> or <judge>No</judge> tags.\n 4. Based on your judgment, follow these strict rules:\n a. If the information is useful AND you now have sufficient information to provide a complete final answer, proceed directly to step 5.\n b. If the information is useful BUT you still need more details, you MUST search again with <search>...</search>.\n c. If the information is not useful, you MUST search again with <search>...</search>. You MUST NOT provide an answer in this case.\n 5. Provide your final answer in <answer>...</answer> tags. The <answer> tag marks the end of the task. After providing the <answer>, you MUST stop and generate no further text.\n\n Question: when did zody's go out of business?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "March 1986" ] }, "style": "rule" }
{ "index": 421, "need_tools_kwargs": true, "question": "when did zody's go out of business?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "March 1986" ] }, "question": "when did zody's go out of business?" } } } }
null
nq
[ { "content": "You are a helpful and harmless assistant.", "role": "system" }, { "content": "Answer the given question step by step.\n Instructions:\n 1. First, conduct reasoning inside <think> and </think> tags whenever you receive new information.\n 2. If you need external knowledge, you can search using <search>query</search>.\n 3. When you receive search results, evaluate their usefulness and put your judgment inside <judge>Yes</judge> or <judge>No</judge> tags.\n 4. Based on your judgment, follow these strict rules:\n a. If the information is useful AND you now have sufficient information to provide a complete final answer, proceed directly to step 5.\n b. If the information is useful BUT you still need more details, you MUST search again with <search>...</search>.\n c. If the information is not useful, you MUST search again with <search>...</search>. You MUST NOT provide an answer in this case.\n 5. Provide your final answer in <answer>...</answer> tags. The <answer> tag marks the end of the task. After providing the <answer>, you MUST stop and generate no further text.\n\n Question: who sings malcom in the middle theme song?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "They Might Be Giants" ] }, "style": "rule" }
{ "index": 422, "need_tools_kwargs": true, "question": "who sings malcom in the middle theme song?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "They Might Be Giants" ] }, "question": "who sings malcom in the middle theme song?" } } } }
null
nq
[ { "content": "You are a helpful and harmless assistant.", "role": "system" }, { "content": "Answer the given question step by step.\n Instructions:\n 1. First, conduct reasoning inside <think> and </think> tags whenever you receive new information.\n 2. If you need external knowledge, you can search using <search>query</search>.\n 3. When you receive search results, evaluate their usefulness and put your judgment inside <judge>Yes</judge> or <judge>No</judge> tags.\n 4. Based on your judgment, follow these strict rules:\n a. If the information is useful AND you now have sufficient information to provide a complete final answer, proceed directly to step 5.\n b. If the information is useful BUT you still need more details, you MUST search again with <search>...</search>.\n c. If the information is not useful, you MUST search again with <search>...</search>. You MUST NOT provide an answer in this case.\n 5. Provide your final answer in <answer>...</answer> tags. The <answer> tag marks the end of the task. After providing the <answer>, you MUST stop and generate no further text.\n\n Question: how big is ganymede compared to earth's moon?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "1.51 times the Moon's" ] }, "style": "rule" }
{ "index": 423, "need_tools_kwargs": true, "question": "how big is ganymede compared to earth's moon?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "1.51 times the Moon's" ] }, "question": "how big is ganymede compared to earth's moon?" } } } }
null
nq
[ { "content": "You are a helpful and harmless assistant.", "role": "system" }, { "content": "Answer the given question step by step.\n Instructions:\n 1. First, conduct reasoning inside <think> and </think> tags whenever you receive new information.\n 2. If you need external knowledge, you can search using <search>query</search>.\n 3. When you receive search results, evaluate their usefulness and put your judgment inside <judge>Yes</judge> or <judge>No</judge> tags.\n 4. Based on your judgment, follow these strict rules:\n a. If the information is useful AND you now have sufficient information to provide a complete final answer, proceed directly to step 5.\n b. If the information is useful BUT you still need more details, you MUST search again with <search>...</search>.\n c. If the information is not useful, you MUST search again with <search>...</search>. You MUST NOT provide an answer in this case.\n 5. Provide your final answer in <answer>...</answer> tags. The <answer> tag marks the end of the task. After providing the <answer>, you MUST stop and generate no further text.\n\n Question: what is the furthest point from the sea in the uk?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "Coton in the Elms" ] }, "style": "rule" }
{ "index": 424, "need_tools_kwargs": true, "question": "what is the furthest point from the sea in the uk?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "Coton in the Elms" ] }, "question": "what is the furthest point from the sea in the uk?" } } } }
null
nq
[ { "content": "You are a helpful and harmless assistant.", "role": "system" }, { "content": "Answer the given question step by step.\n Instructions:\n 1. First, conduct reasoning inside <think> and </think> tags whenever you receive new information.\n 2. If you need external knowledge, you can search using <search>query</search>.\n 3. When you receive search results, evaluate their usefulness and put your judgment inside <judge>Yes</judge> or <judge>No</judge> tags.\n 4. Based on your judgment, follow these strict rules:\n a. If the information is useful AND you now have sufficient information to provide a complete final answer, proceed directly to step 5.\n b. If the information is useful BUT you still need more details, you MUST search again with <search>...</search>.\n c. If the information is not useful, you MUST search again with <search>...</search>. You MUST NOT provide an answer in this case.\n 5. Provide your final answer in <answer>...</answer> tags. The <answer> tag marks the end of the task. After providing the <answer>, you MUST stop and generate no further text.\n\n Question: what disease does nikki have on young and restless?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "breast cancer", "multiple sclerosis", "alcoholism" ] }, "style": "rule" }
{ "index": 425, "need_tools_kwargs": true, "question": "what disease does nikki have on young and restless?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "breast cancer", "multiple sclerosis", "alcoholism" ] }, "question": "what disease does nikki have on young and restless?" } } } }
null
nq
[ { "content": "You are a helpful and harmless assistant.", "role": "system" }, { "content": "Answer the given question step by step.\n Instructions:\n 1. First, conduct reasoning inside <think> and </think> tags whenever you receive new information.\n 2. If you need external knowledge, you can search using <search>query</search>.\n 3. When you receive search results, evaluate their usefulness and put your judgment inside <judge>Yes</judge> or <judge>No</judge> tags.\n 4. Based on your judgment, follow these strict rules:\n a. If the information is useful AND you now have sufficient information to provide a complete final answer, proceed directly to step 5.\n b. If the information is useful BUT you still need more details, you MUST search again with <search>...</search>.\n c. If the information is not useful, you MUST search again with <search>...</search>. You MUST NOT provide an answer in this case.\n 5. Provide your final answer in <answer>...</answer> tags. The <answer> tag marks the end of the task. After providing the <answer>, you MUST stop and generate no further text.\n\n Question: when did they start celebrating cinco de mayo?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "1863" ] }, "style": "rule" }
{ "index": 426, "need_tools_kwargs": true, "question": "when did they start celebrating cinco de mayo?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "1863" ] }, "question": "when did they start celebrating cinco de mayo?" } } } }
null
nq
[ { "content": "You are a helpful and harmless assistant.", "role": "system" }, { "content": "Answer the given question step by step.\n Instructions:\n 1. First, conduct reasoning inside <think> and </think> tags whenever you receive new information.\n 2. If you need external knowledge, you can search using <search>query</search>.\n 3. When you receive search results, evaluate their usefulness and put your judgment inside <judge>Yes</judge> or <judge>No</judge> tags.\n 4. Based on your judgment, follow these strict rules:\n a. If the information is useful AND you now have sufficient information to provide a complete final answer, proceed directly to step 5.\n b. If the information is useful BUT you still need more details, you MUST search again with <search>...</search>.\n c. If the information is not useful, you MUST search again with <search>...</search>. You MUST NOT provide an answer in this case.\n 5. Provide your final answer in <answer>...</answer> tags. The <answer> tag marks the end of the task. After providing the <answer>, you MUST stop and generate no further text.\n\n Question: who was the first governor of western australia?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "Captain Sir James Stirling RN" ] }, "style": "rule" }
{ "index": 427, "need_tools_kwargs": true, "question": "who was the first governor of western australia?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "Captain Sir James Stirling RN" ] }, "question": "who was the first governor of western australia?" } } } }
null
nq
[ { "content": "You are a helpful and harmless assistant.", "role": "system" }, { "content": "Answer the given question step by step.\n Instructions:\n 1. First, conduct reasoning inside <think> and </think> tags whenever you receive new information.\n 2. If you need external knowledge, you can search using <search>query</search>.\n 3. When you receive search results, evaluate their usefulness and put your judgment inside <judge>Yes</judge> or <judge>No</judge> tags.\n 4. Based on your judgment, follow these strict rules:\n a. If the information is useful AND you now have sufficient information to provide a complete final answer, proceed directly to step 5.\n b. If the information is useful BUT you still need more details, you MUST search again with <search>...</search>.\n c. If the information is not useful, you MUST search again with <search>...</search>. You MUST NOT provide an answer in this case.\n 5. Provide your final answer in <answer>...</answer> tags. The <answer> tag marks the end of the task. After providing the <answer>, you MUST stop and generate no further text.\n\n Question: what is the meaning of the name patricia?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "patrician" ] }, "style": "rule" }
{ "index": 428, "need_tools_kwargs": true, "question": "what is the meaning of the name patricia?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "patrician" ] }, "question": "what is the meaning of the name patricia?" } } } }
null
nq
[ { "content": "You are a helpful and harmless assistant.", "role": "system" }, { "content": "Answer the given question step by step.\n Instructions:\n 1. First, conduct reasoning inside <think> and </think> tags whenever you receive new information.\n 2. If you need external knowledge, you can search using <search>query</search>.\n 3. When you receive search results, evaluate their usefulness and put your judgment inside <judge>Yes</judge> or <judge>No</judge> tags.\n 4. Based on your judgment, follow these strict rules:\n a. If the information is useful AND you now have sufficient information to provide a complete final answer, proceed directly to step 5.\n b. If the information is useful BUT you still need more details, you MUST search again with <search>...</search>.\n c. If the information is not useful, you MUST search again with <search>...</search>. You MUST NOT provide an answer in this case.\n 5. Provide your final answer in <answer>...</answer> tags. The <answer> tag marks the end of the task. After providing the <answer>, you MUST stop and generate no further text.\n\n Question: how many days does it take for the moon to rotate around the earth?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "27.323" ] }, "style": "rule" }
{ "index": 429, "need_tools_kwargs": true, "question": "how many days does it take for the moon to rotate around the earth?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "27.323" ] }, "question": "how many days does it take for the moon to rotate around the earth?" } } } }
null
nq
[ { "content": "You are a helpful and harmless assistant.", "role": "system" }, { "content": "Answer the given question step by step.\n Instructions:\n 1. First, conduct reasoning inside <think> and </think> tags whenever you receive new information.\n 2. If you need external knowledge, you can search using <search>query</search>.\n 3. When you receive search results, evaluate their usefulness and put your judgment inside <judge>Yes</judge> or <judge>No</judge> tags.\n 4. Based on your judgment, follow these strict rules:\n a. If the information is useful AND you now have sufficient information to provide a complete final answer, proceed directly to step 5.\n b. If the information is useful BUT you still need more details, you MUST search again with <search>...</search>.\n c. If the information is not useful, you MUST search again with <search>...</search>. You MUST NOT provide an answer in this case.\n 5. Provide your final answer in <answer>...</answer> tags. The <answer> tag marks the end of the task. After providing the <answer>, you MUST stop and generate no further text.\n\n Question: who is in the video don't worry be happy?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "Robin Williams", "Bobby McFerrin", "Bill Irwin" ] }, "style": "rule" }
{ "index": 430, "need_tools_kwargs": true, "question": "who is in the video don't worry be happy?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "Robin Williams", "Bobby McFerrin", "Bill Irwin" ] }, "question": "who is in the video don't worry be happy?" } } } }
null
nq
[ { "content": "You are a helpful and harmless assistant.", "role": "system" }, { "content": "Answer the given question step by step.\n Instructions:\n 1. First, conduct reasoning inside <think> and </think> tags whenever you receive new information.\n 2. If you need external knowledge, you can search using <search>query</search>.\n 3. When you receive search results, evaluate their usefulness and put your judgment inside <judge>Yes</judge> or <judge>No</judge> tags.\n 4. Based on your judgment, follow these strict rules:\n a. If the information is useful AND you now have sufficient information to provide a complete final answer, proceed directly to step 5.\n b. If the information is useful BUT you still need more details, you MUST search again with <search>...</search>.\n c. If the information is not useful, you MUST search again with <search>...</search>. You MUST NOT provide an answer in this case.\n 5. Provide your final answer in <answer>...</answer> tags. The <answer> tag marks the end of the task. After providing the <answer>, you MUST stop and generate no further text.\n\n Question: what bridge does the d train go over?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "Manhattan Bridge" ] }, "style": "rule" }
{ "index": 431, "need_tools_kwargs": true, "question": "what bridge does the d train go over?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "Manhattan Bridge" ] }, "question": "what bridge does the d train go over?" } } } }
null
nq
[ { "content": "You are a helpful and harmless assistant.", "role": "system" }, { "content": "Answer the given question step by step.\n Instructions:\n 1. First, conduct reasoning inside <think> and </think> tags whenever you receive new information.\n 2. If you need external knowledge, you can search using <search>query</search>.\n 3. When you receive search results, evaluate their usefulness and put your judgment inside <judge>Yes</judge> or <judge>No</judge> tags.\n 4. Based on your judgment, follow these strict rules:\n a. If the information is useful AND you now have sufficient information to provide a complete final answer, proceed directly to step 5.\n b. If the information is useful BUT you still need more details, you MUST search again with <search>...</search>.\n c. If the information is not useful, you MUST search again with <search>...</search>. You MUST NOT provide an answer in this case.\n 5. Provide your final answer in <answer>...</answer> tags. The <answer> tag marks the end of the task. After providing the <answer>, you MUST stop and generate no further text.\n\n Question: when did the modern tie come into fashion?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "With the industrial revolution" ] }, "style": "rule" }
{ "index": 432, "need_tools_kwargs": true, "question": "when did the modern tie come into fashion?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "With the industrial revolution" ] }, "question": "when did the modern tie come into fashion?" } } } }
null
nq
[ { "content": "You are a helpful and harmless assistant.", "role": "system" }, { "content": "Answer the given question step by step.\n Instructions:\n 1. First, conduct reasoning inside <think> and </think> tags whenever you receive new information.\n 2. If you need external knowledge, you can search using <search>query</search>.\n 3. When you receive search results, evaluate their usefulness and put your judgment inside <judge>Yes</judge> or <judge>No</judge> tags.\n 4. Based on your judgment, follow these strict rules:\n a. If the information is useful AND you now have sufficient information to provide a complete final answer, proceed directly to step 5.\n b. If the information is useful BUT you still need more details, you MUST search again with <search>...</search>.\n c. If the information is not useful, you MUST search again with <search>...</search>. You MUST NOT provide an answer in this case.\n 5. Provide your final answer in <answer>...</answer> tags. The <answer> tag marks the end of the task. After providing the <answer>, you MUST stop and generate no further text.\n\n Question: where did river niger and river benue meet?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "at Lokoja in Nigeria" ] }, "style": "rule" }
{ "index": 433, "need_tools_kwargs": true, "question": "where did river niger and river benue meet?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "at Lokoja in Nigeria" ] }, "question": "where did river niger and river benue meet?" } } } }
null
nq
[ { "content": "You are a helpful and harmless assistant.", "role": "system" }, { "content": "Answer the given question step by step.\n Instructions:\n 1. First, conduct reasoning inside <think> and </think> tags whenever you receive new information.\n 2. If you need external knowledge, you can search using <search>query</search>.\n 3. When you receive search results, evaluate their usefulness and put your judgment inside <judge>Yes</judge> or <judge>No</judge> tags.\n 4. Based on your judgment, follow these strict rules:\n a. If the information is useful AND you now have sufficient information to provide a complete final answer, proceed directly to step 5.\n b. If the information is useful BUT you still need more details, you MUST search again with <search>...</search>.\n c. If the information is not useful, you MUST search again with <search>...</search>. You MUST NOT provide an answer in this case.\n 5. Provide your final answer in <answer>...</answer> tags. The <answer> tag marks the end of the task. After providing the <answer>, you MUST stop and generate no further text.\n\n Question: who sang i think it so groovy now?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "Friend & Lover" ] }, "style": "rule" }
{ "index": 434, "need_tools_kwargs": true, "question": "who sang i think it so groovy now?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "Friend & Lover" ] }, "question": "who sang i think it so groovy now?" } } } }
null
nq
[ { "content": "You are a helpful and harmless assistant.", "role": "system" }, { "content": "Answer the given question step by step.\n Instructions:\n 1. First, conduct reasoning inside <think> and </think> tags whenever you receive new information.\n 2. If you need external knowledge, you can search using <search>query</search>.\n 3. When you receive search results, evaluate their usefulness and put your judgment inside <judge>Yes</judge> or <judge>No</judge> tags.\n 4. Based on your judgment, follow these strict rules:\n a. If the information is useful AND you now have sufficient information to provide a complete final answer, proceed directly to step 5.\n b. If the information is useful BUT you still need more details, you MUST search again with <search>...</search>.\n c. If the information is not useful, you MUST search again with <search>...</search>. You MUST NOT provide an answer in this case.\n 5. Provide your final answer in <answer>...</answer> tags. The <answer> tag marks the end of the task. After providing the <answer>, you MUST stop and generate no further text.\n\n Question: who was the first black student admitted to the university of mississippi?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "James Howard Meredith" ] }, "style": "rule" }
{ "index": 435, "need_tools_kwargs": true, "question": "who was the first black student admitted to the university of mississippi?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "James Howard Meredith" ] }, "question": "who was the first black student admitted to the university of mississippi?" } } } }
null
nq
[ { "content": "You are a helpful and harmless assistant.", "role": "system" }, { "content": "Answer the given question step by step.\n Instructions:\n 1. First, conduct reasoning inside <think> and </think> tags whenever you receive new information.\n 2. If you need external knowledge, you can search using <search>query</search>.\n 3. When you receive search results, evaluate their usefulness and put your judgment inside <judge>Yes</judge> or <judge>No</judge> tags.\n 4. Based on your judgment, follow these strict rules:\n a. If the information is useful AND you now have sufficient information to provide a complete final answer, proceed directly to step 5.\n b. If the information is useful BUT you still need more details, you MUST search again with <search>...</search>.\n c. If the information is not useful, you MUST search again with <search>...</search>. You MUST NOT provide an answer in this case.\n 5. Provide your final answer in <answer>...</answer> tags. The <answer> tag marks the end of the task. After providing the <answer>, you MUST stop and generate no further text.\n\n Question: where is the capital city located in california?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "Sacramento" ] }, "style": "rule" }
{ "index": 436, "need_tools_kwargs": true, "question": "where is the capital city located in california?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "Sacramento" ] }, "question": "where is the capital city located in california?" } } } }
null
nq
[ { "content": "You are a helpful and harmless assistant.", "role": "system" }, { "content": "Answer the given question step by step.\n Instructions:\n 1. First, conduct reasoning inside <think> and </think> tags whenever you receive new information.\n 2. If you need external knowledge, you can search using <search>query</search>.\n 3. When you receive search results, evaluate their usefulness and put your judgment inside <judge>Yes</judge> or <judge>No</judge> tags.\n 4. Based on your judgment, follow these strict rules:\n a. If the information is useful AND you now have sufficient information to provide a complete final answer, proceed directly to step 5.\n b. If the information is useful BUT you still need more details, you MUST search again with <search>...</search>.\n c. If the information is not useful, you MUST search again with <search>...</search>. You MUST NOT provide an answer in this case.\n 5. Provide your final answer in <answer>...</answer> tags. The <answer> tag marks the end of the task. After providing the <answer>, you MUST stop and generate no further text.\n\n Question: who does the voice of ralph in wreck it ralph?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "John C. Reilly" ] }, "style": "rule" }
{ "index": 437, "need_tools_kwargs": true, "question": "who does the voice of ralph in wreck it ralph?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "John C. Reilly" ] }, "question": "who does the voice of ralph in wreck it ralph?" } } } }
null
nq
[ { "content": "You are a helpful and harmless assistant.", "role": "system" }, { "content": "Answer the given question step by step.\n Instructions:\n 1. First, conduct reasoning inside <think> and </think> tags whenever you receive new information.\n 2. If you need external knowledge, you can search using <search>query</search>.\n 3. When you receive search results, evaluate their usefulness and put your judgment inside <judge>Yes</judge> or <judge>No</judge> tags.\n 4. Based on your judgment, follow these strict rules:\n a. If the information is useful AND you now have sufficient information to provide a complete final answer, proceed directly to step 5.\n b. If the information is useful BUT you still need more details, you MUST search again with <search>...</search>.\n c. If the information is not useful, you MUST search again with <search>...</search>. You MUST NOT provide an answer in this case.\n 5. Provide your final answer in <answer>...</answer> tags. The <answer> tag marks the end of the task. After providing the <answer>, you MUST stop and generate no further text.\n\n Question: where do they live in cheaper by the dozen?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "Evanston, Illinois" ] }, "style": "rule" }
{ "index": 438, "need_tools_kwargs": true, "question": "where do they live in cheaper by the dozen?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "Evanston, Illinois" ] }, "question": "where do they live in cheaper by the dozen?" } } } }
null
nq
[ { "content": "You are a helpful and harmless assistant.", "role": "system" }, { "content": "Answer the given question step by step.\n Instructions:\n 1. First, conduct reasoning inside <think> and </think> tags whenever you receive new information.\n 2. If you need external knowledge, you can search using <search>query</search>.\n 3. When you receive search results, evaluate their usefulness and put your judgment inside <judge>Yes</judge> or <judge>No</judge> tags.\n 4. Based on your judgment, follow these strict rules:\n a. If the information is useful AND you now have sufficient information to provide a complete final answer, proceed directly to step 5.\n b. If the information is useful BUT you still need more details, you MUST search again with <search>...</search>.\n c. If the information is not useful, you MUST search again with <search>...</search>. You MUST NOT provide an answer in this case.\n 5. Provide your final answer in <answer>...</answer> tags. The <answer> tag marks the end of the task. After providing the <answer>, you MUST stop and generate no further text.\n\n Question: where is willy wonka and the chocolate factory based?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "an unnamed town" ] }, "style": "rule" }
{ "index": 439, "need_tools_kwargs": true, "question": "where is willy wonka and the chocolate factory based?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "an unnamed town" ] }, "question": "where is willy wonka and the chocolate factory based?" } } } }
null
nq
[ { "content": "You are a helpful and harmless assistant.", "role": "system" }, { "content": "Answer the given question step by step.\n Instructions:\n 1. First, conduct reasoning inside <think> and </think> tags whenever you receive new information.\n 2. If you need external knowledge, you can search using <search>query</search>.\n 3. When you receive search results, evaluate their usefulness and put your judgment inside <judge>Yes</judge> or <judge>No</judge> tags.\n 4. Based on your judgment, follow these strict rules:\n a. If the information is useful AND you now have sufficient information to provide a complete final answer, proceed directly to step 5.\n b. If the information is useful BUT you still need more details, you MUST search again with <search>...</search>.\n c. If the information is not useful, you MUST search again with <search>...</search>. You MUST NOT provide an answer in this case.\n 5. Provide your final answer in <answer>...</answer> tags. The <answer> tag marks the end of the task. After providing the <answer>, you MUST stop and generate no further text.\n\n Question: real name of raj chandra in rani rashmoni?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "Babughat" ] }, "style": "rule" }
{ "index": 440, "need_tools_kwargs": true, "question": "real name of raj chandra in rani rashmoni?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "Babughat" ] }, "question": "real name of raj chandra in rani rashmoni?" } } } }
null
nq
[ { "content": "You are a helpful and harmless assistant.", "role": "system" }, { "content": "Answer the given question step by step.\n Instructions:\n 1. First, conduct reasoning inside <think> and </think> tags whenever you receive new information.\n 2. If you need external knowledge, you can search using <search>query</search>.\n 3. When you receive search results, evaluate their usefulness and put your judgment inside <judge>Yes</judge> or <judge>No</judge> tags.\n 4. Based on your judgment, follow these strict rules:\n a. If the information is useful AND you now have sufficient information to provide a complete final answer, proceed directly to step 5.\n b. If the information is useful BUT you still need more details, you MUST search again with <search>...</search>.\n c. If the information is not useful, you MUST search again with <search>...</search>. You MUST NOT provide an answer in this case.\n 5. Provide your final answer in <answer>...</answer> tags. The <answer> tag marks the end of the task. After providing the <answer>, you MUST stop and generate no further text.\n\n Question: top scorer of uefa champions league of all time?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "Cristiano Ronaldo" ] }, "style": "rule" }
{ "index": 441, "need_tools_kwargs": true, "question": "top scorer of uefa champions league of all time?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "Cristiano Ronaldo" ] }, "question": "top scorer of uefa champions league of all time?" } } } }
null
nq
[ { "content": "You are a helpful and harmless assistant.", "role": "system" }, { "content": "Answer the given question step by step.\n Instructions:\n 1. First, conduct reasoning inside <think> and </think> tags whenever you receive new information.\n 2. If you need external knowledge, you can search using <search>query</search>.\n 3. When you receive search results, evaluate their usefulness and put your judgment inside <judge>Yes</judge> or <judge>No</judge> tags.\n 4. Based on your judgment, follow these strict rules:\n a. If the information is useful AND you now have sufficient information to provide a complete final answer, proceed directly to step 5.\n b. If the information is useful BUT you still need more details, you MUST search again with <search>...</search>.\n c. If the information is not useful, you MUST search again with <search>...</search>. You MUST NOT provide an answer in this case.\n 5. Provide your final answer in <answer>...</answer> tags. The <answer> tag marks the end of the task. After providing the <answer>, you MUST stop and generate no further text.\n\n Question: who gets married in the last episode of the office?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "Angela Martin", "Dwight Schrute" ] }, "style": "rule" }
{ "index": 442, "need_tools_kwargs": true, "question": "who gets married in the last episode of the office?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "Angela Martin", "Dwight Schrute" ] }, "question": "who gets married in the last episode of the office?" } } } }
null
nq
[ { "content": "You are a helpful and harmless assistant.", "role": "system" }, { "content": "Answer the given question step by step.\n Instructions:\n 1. First, conduct reasoning inside <think> and </think> tags whenever you receive new information.\n 2. If you need external knowledge, you can search using <search>query</search>.\n 3. When you receive search results, evaluate their usefulness and put your judgment inside <judge>Yes</judge> or <judge>No</judge> tags.\n 4. Based on your judgment, follow these strict rules:\n a. If the information is useful AND you now have sufficient information to provide a complete final answer, proceed directly to step 5.\n b. If the information is useful BUT you still need more details, you MUST search again with <search>...</search>.\n c. If the information is not useful, you MUST search again with <search>...</search>. You MUST NOT provide an answer in this case.\n 5. Provide your final answer in <answer>...</answer> tags. The <answer> tag marks the end of the task. After providing the <answer>, you MUST stop and generate no further text.\n\n Question: what is the most threatened animal in the world?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "Actinote zikani" ] }, "style": "rule" }
{ "index": 443, "need_tools_kwargs": true, "question": "what is the most threatened animal in the world?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "Actinote zikani" ] }, "question": "what is the most threatened animal in the world?" } } } }
null
nq
[ { "content": "You are a helpful and harmless assistant.", "role": "system" }, { "content": "Answer the given question step by step.\n Instructions:\n 1. First, conduct reasoning inside <think> and </think> tags whenever you receive new information.\n 2. If you need external knowledge, you can search using <search>query</search>.\n 3. When you receive search results, evaluate their usefulness and put your judgment inside <judge>Yes</judge> or <judge>No</judge> tags.\n 4. Based on your judgment, follow these strict rules:\n a. If the information is useful AND you now have sufficient information to provide a complete final answer, proceed directly to step 5.\n b. If the information is useful BUT you still need more details, you MUST search again with <search>...</search>.\n c. If the information is not useful, you MUST search again with <search>...</search>. You MUST NOT provide an answer in this case.\n 5. Provide your final answer in <answer>...</answer> tags. The <answer> tag marks the end of the task. After providing the <answer>, you MUST stop and generate no further text.\n\n Question: when is the last time the browns won?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "2016" ] }, "style": "rule" }
{ "index": 444, "need_tools_kwargs": true, "question": "when is the last time the browns won?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "2016" ] }, "question": "when is the last time the browns won?" } } } }
null
nq
[ { "content": "You are a helpful and harmless assistant.", "role": "system" }, { "content": "Answer the given question step by step.\n Instructions:\n 1. First, conduct reasoning inside <think> and </think> tags whenever you receive new information.\n 2. If you need external knowledge, you can search using <search>query</search>.\n 3. When you receive search results, evaluate their usefulness and put your judgment inside <judge>Yes</judge> or <judge>No</judge> tags.\n 4. Based on your judgment, follow these strict rules:\n a. If the information is useful AND you now have sufficient information to provide a complete final answer, proceed directly to step 5.\n b. If the information is useful BUT you still need more details, you MUST search again with <search>...</search>.\n c. If the information is not useful, you MUST search again with <search>...</search>. You MUST NOT provide an answer in this case.\n 5. Provide your final answer in <answer>...</answer> tags. The <answer> tag marks the end of the task. After providing the <answer>, you MUST stop and generate no further text.\n\n Question: how many asian countries qualified in world cup 2018?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "5" ] }, "style": "rule" }
{ "index": 445, "need_tools_kwargs": true, "question": "how many asian countries qualified in world cup 2018?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "5" ] }, "question": "how many asian countries qualified in world cup 2018?" } } } }
null
nq
[ { "content": "You are a helpful and harmless assistant.", "role": "system" }, { "content": "Answer the given question step by step.\n Instructions:\n 1. First, conduct reasoning inside <think> and </think> tags whenever you receive new information.\n 2. If you need external knowledge, you can search using <search>query</search>.\n 3. When you receive search results, evaluate their usefulness and put your judgment inside <judge>Yes</judge> or <judge>No</judge> tags.\n 4. Based on your judgment, follow these strict rules:\n a. If the information is useful AND you now have sufficient information to provide a complete final answer, proceed directly to step 5.\n b. If the information is useful BUT you still need more details, you MUST search again with <search>...</search>.\n c. If the information is not useful, you MUST search again with <search>...</search>. You MUST NOT provide an answer in this case.\n 5. Provide your final answer in <answer>...</answer> tags. The <answer> tag marks the end of the task. After providing the <answer>, you MUST stop and generate no further text.\n\n Question: who has the largest sports contract in the world?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "Giancarlo Stanton" ] }, "style": "rule" }
{ "index": 446, "need_tools_kwargs": true, "question": "who has the largest sports contract in the world?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "Giancarlo Stanton" ] }, "question": "who has the largest sports contract in the world?" } } } }
null
nq
[ { "content": "You are a helpful and harmless assistant.", "role": "system" }, { "content": "Answer the given question step by step.\n Instructions:\n 1. First, conduct reasoning inside <think> and </think> tags whenever you receive new information.\n 2. If you need external knowledge, you can search using <search>query</search>.\n 3. When you receive search results, evaluate their usefulness and put your judgment inside <judge>Yes</judge> or <judge>No</judge> tags.\n 4. Based on your judgment, follow these strict rules:\n a. If the information is useful AND you now have sufficient information to provide a complete final answer, proceed directly to step 5.\n b. If the information is useful BUT you still need more details, you MUST search again with <search>...</search>.\n c. If the information is not useful, you MUST search again with <search>...</search>. You MUST NOT provide an answer in this case.\n 5. Provide your final answer in <answer>...</answer> tags. The <answer> tag marks the end of the task. After providing the <answer>, you MUST stop and generate no further text.\n\n Question: who came up with the idea of communism?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "Karl Marx" ] }, "style": "rule" }
{ "index": 447, "need_tools_kwargs": true, "question": "who came up with the idea of communism?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "Karl Marx" ] }, "question": "who came up with the idea of communism?" } } } }
null
nq
[ { "content": "You are a helpful and harmless assistant.", "role": "system" }, { "content": "Answer the given question step by step.\n Instructions:\n 1. First, conduct reasoning inside <think> and </think> tags whenever you receive new information.\n 2. If you need external knowledge, you can search using <search>query</search>.\n 3. When you receive search results, evaluate their usefulness and put your judgment inside <judge>Yes</judge> or <judge>No</judge> tags.\n 4. Based on your judgment, follow these strict rules:\n a. If the information is useful AND you now have sufficient information to provide a complete final answer, proceed directly to step 5.\n b. If the information is useful BUT you still need more details, you MUST search again with <search>...</search>.\n c. If the information is not useful, you MUST search again with <search>...</search>. You MUST NOT provide an answer in this case.\n 5. Provide your final answer in <answer>...</answer> tags. The <answer> tag marks the end of the task. After providing the <answer>, you MUST stop and generate no further text.\n\n Question: what did hal stand for in 2001 a space odyssey?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "Heuristically programmed ALgorithmic" ] }, "style": "rule" }
{ "index": 448, "need_tools_kwargs": true, "question": "what did hal stand for in 2001 a space odyssey?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "Heuristically programmed ALgorithmic" ] }, "question": "what did hal stand for in 2001 a space odyssey?" } } } }
null
nq
[ { "content": "You are a helpful and harmless assistant.", "role": "system" }, { "content": "Answer the given question step by step.\n Instructions:\n 1. First, conduct reasoning inside <think> and </think> tags whenever you receive new information.\n 2. If you need external knowledge, you can search using <search>query</search>.\n 3. When you receive search results, evaluate their usefulness and put your judgment inside <judge>Yes</judge> or <judge>No</judge> tags.\n 4. Based on your judgment, follow these strict rules:\n a. If the information is useful AND you now have sufficient information to provide a complete final answer, proceed directly to step 5.\n b. If the information is useful BUT you still need more details, you MUST search again with <search>...</search>.\n c. If the information is not useful, you MUST search again with <search>...</search>. You MUST NOT provide an answer in this case.\n 5. Provide your final answer in <answer>...</answer> tags. The <answer> tag marks the end of the task. After providing the <answer>, you MUST stop and generate no further text.\n\n Question: who wrote the music for how the grinch stole christmas?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "Dr. Seuss" ] }, "style": "rule" }
{ "index": 449, "need_tools_kwargs": true, "question": "who wrote the music for how the grinch stole christmas?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "Dr. Seuss" ] }, "question": "who wrote the music for how the grinch stole christmas?" } } } }
null
nq
[ { "content": "You are a helpful and harmless assistant.", "role": "system" }, { "content": "Answer the given question step by step.\n Instructions:\n 1. First, conduct reasoning inside <think> and </think> tags whenever you receive new information.\n 2. If you need external knowledge, you can search using <search>query</search>.\n 3. When you receive search results, evaluate their usefulness and put your judgment inside <judge>Yes</judge> or <judge>No</judge> tags.\n 4. Based on your judgment, follow these strict rules:\n a. If the information is useful AND you now have sufficient information to provide a complete final answer, proceed directly to step 5.\n b. If the information is useful BUT you still need more details, you MUST search again with <search>...</search>.\n c. If the information is not useful, you MUST search again with <search>...</search>. You MUST NOT provide an answer in this case.\n 5. Provide your final answer in <answer>...</answer> tags. The <answer> tag marks the end of the task. After providing the <answer>, you MUST stop and generate no further text.\n\n Question: who has to push the rock up the hill?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "Sisyphus" ] }, "style": "rule" }
{ "index": 450, "need_tools_kwargs": true, "question": "who has to push the rock up the hill?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "Sisyphus" ] }, "question": "who has to push the rock up the hill?" } } } }
null
nq
[ { "content": "You are a helpful and harmless assistant.", "role": "system" }, { "content": "Answer the given question step by step.\n Instructions:\n 1. First, conduct reasoning inside <think> and </think> tags whenever you receive new information.\n 2. If you need external knowledge, you can search using <search>query</search>.\n 3. When you receive search results, evaluate their usefulness and put your judgment inside <judge>Yes</judge> or <judge>No</judge> tags.\n 4. Based on your judgment, follow these strict rules:\n a. If the information is useful AND you now have sufficient information to provide a complete final answer, proceed directly to step 5.\n b. If the information is useful BUT you still need more details, you MUST search again with <search>...</search>.\n c. If the information is not useful, you MUST search again with <search>...</search>. You MUST NOT provide an answer in this case.\n 5. Provide your final answer in <answer>...</answer> tags. The <answer> tag marks the end of the task. After providing the <answer>, you MUST stop and generate no further text.\n\n Question: when did the second fleet arrive in australia?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "28 June 1790", "20 June 1790", "27 June 1790", "3 June 1790", "26 June 1790" ] }, "style": "rule" }
{ "index": 451, "need_tools_kwargs": true, "question": "when did the second fleet arrive in australia?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "28 June 1790", "20 June 1790", "27 June 1790", "3 June 1790", "26 June 1790" ] }, "question": "when did the second fleet arrive in australia?" } } } }
null
nq
[ { "content": "You are a helpful and harmless assistant.", "role": "system" }, { "content": "Answer the given question step by step.\n Instructions:\n 1. First, conduct reasoning inside <think> and </think> tags whenever you receive new information.\n 2. If you need external knowledge, you can search using <search>query</search>.\n 3. When you receive search results, evaluate their usefulness and put your judgment inside <judge>Yes</judge> or <judge>No</judge> tags.\n 4. Based on your judgment, follow these strict rules:\n a. If the information is useful AND you now have sufficient information to provide a complete final answer, proceed directly to step 5.\n b. If the information is useful BUT you still need more details, you MUST search again with <search>...</search>.\n c. If the information is not useful, you MUST search again with <search>...</search>. You MUST NOT provide an answer in this case.\n 5. Provide your final answer in <answer>...</answer> tags. The <answer> tag marks the end of the task. After providing the <answer>, you MUST stop and generate no further text.\n\n Question: scottish architect who developed st martins in the field?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "James Gibbs" ] }, "style": "rule" }
{ "index": 452, "need_tools_kwargs": true, "question": "scottish architect who developed st martins in the field?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "James Gibbs" ] }, "question": "scottish architect who developed st martins in the field?" } } } }
null
nq
[ { "content": "You are a helpful and harmless assistant.", "role": "system" }, { "content": "Answer the given question step by step.\n Instructions:\n 1. First, conduct reasoning inside <think> and </think> tags whenever you receive new information.\n 2. If you need external knowledge, you can search using <search>query</search>.\n 3. When you receive search results, evaluate their usefulness and put your judgment inside <judge>Yes</judge> or <judge>No</judge> tags.\n 4. Based on your judgment, follow these strict rules:\n a. If the information is useful AND you now have sufficient information to provide a complete final answer, proceed directly to step 5.\n b. If the information is useful BUT you still need more details, you MUST search again with <search>...</search>.\n c. If the information is not useful, you MUST search again with <search>...</search>. You MUST NOT provide an answer in this case.\n 5. Provide your final answer in <answer>...</answer> tags. The <answer> tag marks the end of the task. After providing the <answer>, you MUST stop and generate no further text.\n\n Question: who played king kong in kong skull island?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "Toby Kebbell", "Terry Notary" ] }, "style": "rule" }
{ "index": 453, "need_tools_kwargs": true, "question": "who played king kong in kong skull island?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "Toby Kebbell", "Terry Notary" ] }, "question": "who played king kong in kong skull island?" } } } }
null
nq
[ { "content": "You are a helpful and harmless assistant.", "role": "system" }, { "content": "Answer the given question step by step.\n Instructions:\n 1. First, conduct reasoning inside <think> and </think> tags whenever you receive new information.\n 2. If you need external knowledge, you can search using <search>query</search>.\n 3. When you receive search results, evaluate their usefulness and put your judgment inside <judge>Yes</judge> or <judge>No</judge> tags.\n 4. Based on your judgment, follow these strict rules:\n a. If the information is useful AND you now have sufficient information to provide a complete final answer, proceed directly to step 5.\n b. If the information is useful BUT you still need more details, you MUST search again with <search>...</search>.\n c. If the information is not useful, you MUST search again with <search>...</search>. You MUST NOT provide an answer in this case.\n 5. Provide your final answer in <answer>...</answer> tags. The <answer> tag marks the end of the task. After providing the <answer>, you MUST stop and generate no further text.\n\n Question: what type of business structure does cadbury have?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "importing chocolates" ] }, "style": "rule" }
{ "index": 454, "need_tools_kwargs": true, "question": "what type of business structure does cadbury have?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "importing chocolates" ] }, "question": "what type of business structure does cadbury have?" } } } }
null
nq
[ { "content": "You are a helpful and harmless assistant.", "role": "system" }, { "content": "Answer the given question step by step.\n Instructions:\n 1. First, conduct reasoning inside <think> and </think> tags whenever you receive new information.\n 2. If you need external knowledge, you can search using <search>query</search>.\n 3. When you receive search results, evaluate their usefulness and put your judgment inside <judge>Yes</judge> or <judge>No</judge> tags.\n 4. Based on your judgment, follow these strict rules:\n a. If the information is useful AND you now have sufficient information to provide a complete final answer, proceed directly to step 5.\n b. If the information is useful BUT you still need more details, you MUST search again with <search>...</search>.\n c. If the information is not useful, you MUST search again with <search>...</search>. You MUST NOT provide an answer in this case.\n 5. Provide your final answer in <answer>...</answer> tags. The <answer> tag marks the end of the task. After providing the <answer>, you MUST stop and generate no further text.\n\n Question: when did the american civil rights movement end?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "1968" ] }, "style": "rule" }
{ "index": 455, "need_tools_kwargs": true, "question": "when did the american civil rights movement end?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "1968" ] }, "question": "when did the american civil rights movement end?" } } } }
null
nq
[ { "content": "You are a helpful and harmless assistant.", "role": "system" }, { "content": "Answer the given question step by step.\n Instructions:\n 1. First, conduct reasoning inside <think> and </think> tags whenever you receive new information.\n 2. If you need external knowledge, you can search using <search>query</search>.\n 3. When you receive search results, evaluate their usefulness and put your judgment inside <judge>Yes</judge> or <judge>No</judge> tags.\n 4. Based on your judgment, follow these strict rules:\n a. If the information is useful AND you now have sufficient information to provide a complete final answer, proceed directly to step 5.\n b. If the information is useful BUT you still need more details, you MUST search again with <search>...</search>.\n c. If the information is not useful, you MUST search again with <search>...</search>. You MUST NOT provide an answer in this case.\n 5. Provide your final answer in <answer>...</answer> tags. The <answer> tag marks the end of the task. After providing the <answer>, you MUST stop and generate no further text.\n\n Question: how many episodes will be in 13 reasons why season 2?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "13" ] }, "style": "rule" }
{ "index": 456, "need_tools_kwargs": true, "question": "how many episodes will be in 13 reasons why season 2?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "13" ] }, "question": "how many episodes will be in 13 reasons why season 2?" } } } }
null
nq
[ { "content": "You are a helpful and harmless assistant.", "role": "system" }, { "content": "Answer the given question step by step.\n Instructions:\n 1. First, conduct reasoning inside <think> and </think> tags whenever you receive new information.\n 2. If you need external knowledge, you can search using <search>query</search>.\n 3. When you receive search results, evaluate their usefulness and put your judgment inside <judge>Yes</judge> or <judge>No</judge> tags.\n 4. Based on your judgment, follow these strict rules:\n a. If the information is useful AND you now have sufficient information to provide a complete final answer, proceed directly to step 5.\n b. If the information is useful BUT you still need more details, you MUST search again with <search>...</search>.\n c. If the information is not useful, you MUST search again with <search>...</search>. You MUST NOT provide an answer in this case.\n 5. Provide your final answer in <answer>...</answer> tags. The <answer> tag marks the end of the task. After providing the <answer>, you MUST stop and generate no further text.\n\n Question: what is the current version of php and mysql?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "7.2" ] }, "style": "rule" }
{ "index": 457, "need_tools_kwargs": true, "question": "what is the current version of php and mysql?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "7.2" ] }, "question": "what is the current version of php and mysql?" } } } }
null
nq
[ { "content": "You are a helpful and harmless assistant.", "role": "system" }, { "content": "Answer the given question step by step.\n Instructions:\n 1. First, conduct reasoning inside <think> and </think> tags whenever you receive new information.\n 2. If you need external knowledge, you can search using <search>query</search>.\n 3. When you receive search results, evaluate their usefulness and put your judgment inside <judge>Yes</judge> or <judge>No</judge> tags.\n 4. Based on your judgment, follow these strict rules:\n a. If the information is useful AND you now have sufficient information to provide a complete final answer, proceed directly to step 5.\n b. If the information is useful BUT you still need more details, you MUST search again with <search>...</search>.\n c. If the information is not useful, you MUST search again with <search>...</search>. You MUST NOT provide an answer in this case.\n 5. Provide your final answer in <answer>...</answer> tags. The <answer> tag marks the end of the task. After providing the <answer>, you MUST stop and generate no further text.\n\n Question: who is the minister of sanitation in ghana?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "Joseph Kofi Adda" ] }, "style": "rule" }
{ "index": 458, "need_tools_kwargs": true, "question": "who is the minister of sanitation in ghana?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "Joseph Kofi Adda" ] }, "question": "who is the minister of sanitation in ghana?" } } } }
null
nq
[ { "content": "You are a helpful and harmless assistant.", "role": "system" }, { "content": "Answer the given question step by step.\n Instructions:\n 1. First, conduct reasoning inside <think> and </think> tags whenever you receive new information.\n 2. If you need external knowledge, you can search using <search>query</search>.\n 3. When you receive search results, evaluate their usefulness and put your judgment inside <judge>Yes</judge> or <judge>No</judge> tags.\n 4. Based on your judgment, follow these strict rules:\n a. If the information is useful AND you now have sufficient information to provide a complete final answer, proceed directly to step 5.\n b. If the information is useful BUT you still need more details, you MUST search again with <search>...</search>.\n c. If the information is not useful, you MUST search again with <search>...</search>. You MUST NOT provide an answer in this case.\n 5. Provide your final answer in <answer>...</answer> tags. The <answer> tag marks the end of the task. After providing the <answer>, you MUST stop and generate no further text.\n\n Question: when did dade county become miami dade county?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "November 13, 1997" ] }, "style": "rule" }
{ "index": 459, "need_tools_kwargs": true, "question": "when did dade county become miami dade county?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "November 13, 1997" ] }, "question": "when did dade county become miami dade county?" } } } }
null
nq
[ { "content": "You are a helpful and harmless assistant.", "role": "system" }, { "content": "Answer the given question step by step.\n Instructions:\n 1. First, conduct reasoning inside <think> and </think> tags whenever you receive new information.\n 2. If you need external knowledge, you can search using <search>query</search>.\n 3. When you receive search results, evaluate their usefulness and put your judgment inside <judge>Yes</judge> or <judge>No</judge> tags.\n 4. Based on your judgment, follow these strict rules:\n a. If the information is useful AND you now have sufficient information to provide a complete final answer, proceed directly to step 5.\n b. If the information is useful BUT you still need more details, you MUST search again with <search>...</search>.\n c. If the information is not useful, you MUST search again with <search>...</search>. You MUST NOT provide an answer in this case.\n 5. Provide your final answer in <answer>...</answer> tags. The <answer> tag marks the end of the task. After providing the <answer>, you MUST stop and generate no further text.\n\n Question: how many nfl championships did the packers win?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "13" ] }, "style": "rule" }
{ "index": 460, "need_tools_kwargs": true, "question": "how many nfl championships did the packers win?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "13" ] }, "question": "how many nfl championships did the packers win?" } } } }
null
nq
[ { "content": "You are a helpful and harmless assistant.", "role": "system" }, { "content": "Answer the given question step by step.\n Instructions:\n 1. First, conduct reasoning inside <think> and </think> tags whenever you receive new information.\n 2. If you need external knowledge, you can search using <search>query</search>.\n 3. When you receive search results, evaluate their usefulness and put your judgment inside <judge>Yes</judge> or <judge>No</judge> tags.\n 4. Based on your judgment, follow these strict rules:\n a. If the information is useful AND you now have sufficient information to provide a complete final answer, proceed directly to step 5.\n b. If the information is useful BUT you still need more details, you MUST search again with <search>...</search>.\n c. If the information is not useful, you MUST search again with <search>...</search>. You MUST NOT provide an answer in this case.\n 5. Provide your final answer in <answer>...</answer> tags. The <answer> tag marks the end of the task. After providing the <answer>, you MUST stop and generate no further text.\n\n Question: what is an example of a sutural bone?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "Inca bone" ] }, "style": "rule" }
{ "index": 461, "need_tools_kwargs": true, "question": "what is an example of a sutural bone?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "Inca bone" ] }, "question": "what is an example of a sutural bone?" } } } }
null
nq
[ { "content": "You are a helpful and harmless assistant.", "role": "system" }, { "content": "Answer the given question step by step.\n Instructions:\n 1. First, conduct reasoning inside <think> and </think> tags whenever you receive new information.\n 2. If you need external knowledge, you can search using <search>query</search>.\n 3. When you receive search results, evaluate their usefulness and put your judgment inside <judge>Yes</judge> or <judge>No</judge> tags.\n 4. Based on your judgment, follow these strict rules:\n a. If the information is useful AND you now have sufficient information to provide a complete final answer, proceed directly to step 5.\n b. If the information is useful BUT you still need more details, you MUST search again with <search>...</search>.\n c. If the information is not useful, you MUST search again with <search>...</search>. You MUST NOT provide an answer in this case.\n 5. Provide your final answer in <answer>...</answer> tags. The <answer> tag marks the end of the task. After providing the <answer>, you MUST stop and generate no further text.\n\n Question: who becomes the president after the president is impeached?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "Vice President" ] }, "style": "rule" }
{ "index": 462, "need_tools_kwargs": true, "question": "who becomes the president after the president is impeached?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "Vice President" ] }, "question": "who becomes the president after the president is impeached?" } } } }
null
nq
[ { "content": "You are a helpful and harmless assistant.", "role": "system" }, { "content": "Answer the given question step by step.\n Instructions:\n 1. First, conduct reasoning inside <think> and </think> tags whenever you receive new information.\n 2. If you need external knowledge, you can search using <search>query</search>.\n 3. When you receive search results, evaluate their usefulness and put your judgment inside <judge>Yes</judge> or <judge>No</judge> tags.\n 4. Based on your judgment, follow these strict rules:\n a. If the information is useful AND you now have sufficient information to provide a complete final answer, proceed directly to step 5.\n b. If the information is useful BUT you still need more details, you MUST search again with <search>...</search>.\n c. If the information is not useful, you MUST search again with <search>...</search>. You MUST NOT provide an answer in this case.\n 5. Provide your final answer in <answer>...</answer> tags. The <answer> tag marks the end of the task. After providing the <answer>, you MUST stop and generate no further text.\n\n Question: who plays the mexican grandma in jack and jill?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "Eugenio Derbez" ] }, "style": "rule" }
{ "index": 463, "need_tools_kwargs": true, "question": "who plays the mexican grandma in jack and jill?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "Eugenio Derbez" ] }, "question": "who plays the mexican grandma in jack and jill?" } } } }
null
nq
[ { "content": "You are a helpful and harmless assistant.", "role": "system" }, { "content": "Answer the given question step by step.\n Instructions:\n 1. First, conduct reasoning inside <think> and </think> tags whenever you receive new information.\n 2. If you need external knowledge, you can search using <search>query</search>.\n 3. When you receive search results, evaluate their usefulness and put your judgment inside <judge>Yes</judge> or <judge>No</judge> tags.\n 4. Based on your judgment, follow these strict rules:\n a. If the information is useful AND you now have sufficient information to provide a complete final answer, proceed directly to step 5.\n b. If the information is useful BUT you still need more details, you MUST search again with <search>...</search>.\n c. If the information is not useful, you MUST search again with <search>...</search>. You MUST NOT provide an answer in this case.\n 5. Provide your final answer in <answer>...</answer> tags. The <answer> tag marks the end of the task. After providing the <answer>, you MUST stop and generate no further text.\n\n Question: who plays madison grey on hawaii five o?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "Elisabeth Röhm" ] }, "style": "rule" }
{ "index": 464, "need_tools_kwargs": true, "question": "who plays madison grey on hawaii five o?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "Elisabeth Röhm" ] }, "question": "who plays madison grey on hawaii five o?" } } } }
null
nq
[ { "content": "You are a helpful and harmless assistant.", "role": "system" }, { "content": "Answer the given question step by step.\n Instructions:\n 1. First, conduct reasoning inside <think> and </think> tags whenever you receive new information.\n 2. If you need external knowledge, you can search using <search>query</search>.\n 3. When you receive search results, evaluate their usefulness and put your judgment inside <judge>Yes</judge> or <judge>No</judge> tags.\n 4. Based on your judgment, follow these strict rules:\n a. If the information is useful AND you now have sufficient information to provide a complete final answer, proceed directly to step 5.\n b. If the information is useful BUT you still need more details, you MUST search again with <search>...</search>.\n c. If the information is not useful, you MUST search again with <search>...</search>. You MUST NOT provide an answer in this case.\n 5. Provide your final answer in <answer>...</answer> tags. The <answer> tag marks the end of the task. After providing the <answer>, you MUST stop and generate no further text.\n\n Question: when is the new barbie movie coming out?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "May 8, 2020" ] }, "style": "rule" }
{ "index": 465, "need_tools_kwargs": true, "question": "when is the new barbie movie coming out?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "May 8, 2020" ] }, "question": "when is the new barbie movie coming out?" } } } }
null
nq
[ { "content": "You are a helpful and harmless assistant.", "role": "system" }, { "content": "Answer the given question step by step.\n Instructions:\n 1. First, conduct reasoning inside <think> and </think> tags whenever you receive new information.\n 2. If you need external knowledge, you can search using <search>query</search>.\n 3. When you receive search results, evaluate their usefulness and put your judgment inside <judge>Yes</judge> or <judge>No</judge> tags.\n 4. Based on your judgment, follow these strict rules:\n a. If the information is useful AND you now have sufficient information to provide a complete final answer, proceed directly to step 5.\n b. If the information is useful BUT you still need more details, you MUST search again with <search>...</search>.\n c. If the information is not useful, you MUST search again with <search>...</search>. You MUST NOT provide an answer in this case.\n 5. Provide your final answer in <answer>...</answer> tags. The <answer> tag marks the end of the task. After providing the <answer>, you MUST stop and generate no further text.\n\n Question: when were the first documented medical records created?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "19th century" ] }, "style": "rule" }
{ "index": 466, "need_tools_kwargs": true, "question": "when were the first documented medical records created?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "19th century" ] }, "question": "when were the first documented medical records created?" } } } }
null
nq
[ { "content": "You are a helpful and harmless assistant.", "role": "system" }, { "content": "Answer the given question step by step.\n Instructions:\n 1. First, conduct reasoning inside <think> and </think> tags whenever you receive new information.\n 2. If you need external knowledge, you can search using <search>query</search>.\n 3. When you receive search results, evaluate their usefulness and put your judgment inside <judge>Yes</judge> or <judge>No</judge> tags.\n 4. Based on your judgment, follow these strict rules:\n a. If the information is useful AND you now have sufficient information to provide a complete final answer, proceed directly to step 5.\n b. If the information is useful BUT you still need more details, you MUST search again with <search>...</search>.\n c. If the information is not useful, you MUST search again with <search>...</search>. You MUST NOT provide an answer in this case.\n 5. Provide your final answer in <answer>...</answer> tags. The <answer> tag marks the end of the task. After providing the <answer>, you MUST stop and generate no further text.\n\n Question: who plays young gideon in once upon a time?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "Giles Matthey" ] }, "style": "rule" }
{ "index": 467, "need_tools_kwargs": true, "question": "who plays young gideon in once upon a time?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "Giles Matthey" ] }, "question": "who plays young gideon in once upon a time?" } } } }
null
nq
[ { "content": "You are a helpful and harmless assistant.", "role": "system" }, { "content": "Answer the given question step by step.\n Instructions:\n 1. First, conduct reasoning inside <think> and </think> tags whenever you receive new information.\n 2. If you need external knowledge, you can search using <search>query</search>.\n 3. When you receive search results, evaluate their usefulness and put your judgment inside <judge>Yes</judge> or <judge>No</judge> tags.\n 4. Based on your judgment, follow these strict rules:\n a. If the information is useful AND you now have sufficient information to provide a complete final answer, proceed directly to step 5.\n b. If the information is useful BUT you still need more details, you MUST search again with <search>...</search>.\n c. If the information is not useful, you MUST search again with <search>...</search>. You MUST NOT provide an answer in this case.\n 5. Provide your final answer in <answer>...</answer> tags. The <answer> tag marks the end of the task. After providing the <answer>, you MUST stop and generate no further text.\n\n Question: who is the highest paid actor of hollywood?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "Keanu Reeves" ] }, "style": "rule" }
{ "index": 468, "need_tools_kwargs": true, "question": "who is the highest paid actor of hollywood?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "Keanu Reeves" ] }, "question": "who is the highest paid actor of hollywood?" } } } }
null
nq
[ { "content": "You are a helpful and harmless assistant.", "role": "system" }, { "content": "Answer the given question step by step.\n Instructions:\n 1. First, conduct reasoning inside <think> and </think> tags whenever you receive new information.\n 2. If you need external knowledge, you can search using <search>query</search>.\n 3. When you receive search results, evaluate their usefulness and put your judgment inside <judge>Yes</judge> or <judge>No</judge> tags.\n 4. Based on your judgment, follow these strict rules:\n a. If the information is useful AND you now have sufficient information to provide a complete final answer, proceed directly to step 5.\n b. If the information is useful BUT you still need more details, you MUST search again with <search>...</search>.\n c. If the information is not useful, you MUST search again with <search>...</search>. You MUST NOT provide an answer in this case.\n 5. Provide your final answer in <answer>...</answer> tags. The <answer> tag marks the end of the task. After providing the <answer>, you MUST stop and generate no further text.\n\n Question: when does the first friday the 13th take place?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "June 13, 1980" ] }, "style": "rule" }
{ "index": 469, "need_tools_kwargs": true, "question": "when does the first friday the 13th take place?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "June 13, 1980" ] }, "question": "when does the first friday the 13th take place?" } } } }
null
nq
[ { "content": "You are a helpful and harmless assistant.", "role": "system" }, { "content": "Answer the given question step by step.\n Instructions:\n 1. First, conduct reasoning inside <think> and </think> tags whenever you receive new information.\n 2. If you need external knowledge, you can search using <search>query</search>.\n 3. When you receive search results, evaluate their usefulness and put your judgment inside <judge>Yes</judge> or <judge>No</judge> tags.\n 4. Based on your judgment, follow these strict rules:\n a. If the information is useful AND you now have sufficient information to provide a complete final answer, proceed directly to step 5.\n b. If the information is useful BUT you still need more details, you MUST search again with <search>...</search>.\n c. If the information is not useful, you MUST search again with <search>...</search>. You MUST NOT provide an answer in this case.\n 5. Provide your final answer in <answer>...</answer> tags. The <answer> tag marks the end of the task. After providing the <answer>, you MUST stop and generate no further text.\n\n Question: what is the real name of o henry?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "William Sydney Porter" ] }, "style": "rule" }
{ "index": 470, "need_tools_kwargs": true, "question": "what is the real name of o henry?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "William Sydney Porter" ] }, "question": "what is the real name of o henry?" } } } }
null
nq
[ { "content": "You are a helpful and harmless assistant.", "role": "system" }, { "content": "Answer the given question step by step.\n Instructions:\n 1. First, conduct reasoning inside <think> and </think> tags whenever you receive new information.\n 2. If you need external knowledge, you can search using <search>query</search>.\n 3. When you receive search results, evaluate their usefulness and put your judgment inside <judge>Yes</judge> or <judge>No</judge> tags.\n 4. Based on your judgment, follow these strict rules:\n a. If the information is useful AND you now have sufficient information to provide a complete final answer, proceed directly to step 5.\n b. If the information is useful BUT you still need more details, you MUST search again with <search>...</search>.\n c. If the information is not useful, you MUST search again with <search>...</search>. You MUST NOT provide an answer in this case.\n 5. Provide your final answer in <answer>...</answer> tags. The <answer> tag marks the end of the task. After providing the <answer>, you MUST stop and generate no further text.\n\n Question: who sets the salaries for senate and house members?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "Congress" ] }, "style": "rule" }
{ "index": 471, "need_tools_kwargs": true, "question": "who sets the salaries for senate and house members?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "Congress" ] }, "question": "who sets the salaries for senate and house members?" } } } }
null
nq
[ { "content": "You are a helpful and harmless assistant.", "role": "system" }, { "content": "Answer the given question step by step.\n Instructions:\n 1. First, conduct reasoning inside <think> and </think> tags whenever you receive new information.\n 2. If you need external knowledge, you can search using <search>query</search>.\n 3. When you receive search results, evaluate their usefulness and put your judgment inside <judge>Yes</judge> or <judge>No</judge> tags.\n 4. Based on your judgment, follow these strict rules:\n a. If the information is useful AND you now have sufficient information to provide a complete final answer, proceed directly to step 5.\n b. If the information is useful BUT you still need more details, you MUST search again with <search>...</search>.\n c. If the information is not useful, you MUST search again with <search>...</search>. You MUST NOT provide an answer in this case.\n 5. Provide your final answer in <answer>...</answer> tags. The <answer> tag marks the end of the task. After providing the <answer>, you MUST stop and generate no further text.\n\n Question: what is the accent over the i in spanish called?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "acute accent" ] }, "style": "rule" }
{ "index": 472, "need_tools_kwargs": true, "question": "what is the accent over the i in spanish called?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "acute accent" ] }, "question": "what is the accent over the i in spanish called?" } } } }
null
nq
[ { "content": "You are a helpful and harmless assistant.", "role": "system" }, { "content": "Answer the given question step by step.\n Instructions:\n 1. First, conduct reasoning inside <think> and </think> tags whenever you receive new information.\n 2. If you need external knowledge, you can search using <search>query</search>.\n 3. When you receive search results, evaluate their usefulness and put your judgment inside <judge>Yes</judge> or <judge>No</judge> tags.\n 4. Based on your judgment, follow these strict rules:\n a. If the information is useful AND you now have sufficient information to provide a complete final answer, proceed directly to step 5.\n b. If the information is useful BUT you still need more details, you MUST search again with <search>...</search>.\n c. If the information is not useful, you MUST search again with <search>...</search>. You MUST NOT provide an answer in this case.\n 5. Provide your final answer in <answer>...</answer> tags. The <answer> tag marks the end of the task. After providing the <answer>, you MUST stop and generate no further text.\n\n Question: when does lois come to new york in gossip girl?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "\"Petty in Pink\"" ] }, "style": "rule" }
{ "index": 473, "need_tools_kwargs": true, "question": "when does lois come to new york in gossip girl?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "\"Petty in Pink\"" ] }, "question": "when does lois come to new york in gossip girl?" } } } }
null
nq
[ { "content": "You are a helpful and harmless assistant.", "role": "system" }, { "content": "Answer the given question step by step.\n Instructions:\n 1. First, conduct reasoning inside <think> and </think> tags whenever you receive new information.\n 2. If you need external knowledge, you can search using <search>query</search>.\n 3. When you receive search results, evaluate their usefulness and put your judgment inside <judge>Yes</judge> or <judge>No</judge> tags.\n 4. Based on your judgment, follow these strict rules:\n a. If the information is useful AND you now have sufficient information to provide a complete final answer, proceed directly to step 5.\n b. If the information is useful BUT you still need more details, you MUST search again with <search>...</search>.\n c. If the information is not useful, you MUST search again with <search>...</search>. You MUST NOT provide an answer in this case.\n 5. Provide your final answer in <answer>...</answer> tags. The <answer> tag marks the end of the task. After providing the <answer>, you MUST stop and generate no further text.\n\n Question: where did they film ninja warrior in australia?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "Cockatoo Island" ] }, "style": "rule" }
{ "index": 474, "need_tools_kwargs": true, "question": "where did they film ninja warrior in australia?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "Cockatoo Island" ] }, "question": "where did they film ninja warrior in australia?" } } } }
null
nq
[ { "content": "You are a helpful and harmless assistant.", "role": "system" }, { "content": "Answer the given question step by step.\n Instructions:\n 1. First, conduct reasoning inside <think> and </think> tags whenever you receive new information.\n 2. If you need external knowledge, you can search using <search>query</search>.\n 3. When you receive search results, evaluate their usefulness and put your judgment inside <judge>Yes</judge> or <judge>No</judge> tags.\n 4. Based on your judgment, follow these strict rules:\n a. If the information is useful AND you now have sufficient information to provide a complete final answer, proceed directly to step 5.\n b. If the information is useful BUT you still need more details, you MUST search again with <search>...</search>.\n c. If the information is not useful, you MUST search again with <search>...</search>. You MUST NOT provide an answer in this case.\n 5. Provide your final answer in <answer>...</answer> tags. The <answer> tag marks the end of the task. After providing the <answer>, you MUST stop and generate no further text.\n\n Question: what is the full form of icu in hospital?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "intensive care unit" ] }, "style": "rule" }
{ "index": 475, "need_tools_kwargs": true, "question": "what is the full form of icu in hospital?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "intensive care unit" ] }, "question": "what is the full form of icu in hospital?" } } } }
null
nq
[ { "content": "You are a helpful and harmless assistant.", "role": "system" }, { "content": "Answer the given question step by step.\n Instructions:\n 1. First, conduct reasoning inside <think> and </think> tags whenever you receive new information.\n 2. If you need external knowledge, you can search using <search>query</search>.\n 3. When you receive search results, evaluate their usefulness and put your judgment inside <judge>Yes</judge> or <judge>No</judge> tags.\n 4. Based on your judgment, follow these strict rules:\n a. If the information is useful AND you now have sufficient information to provide a complete final answer, proceed directly to step 5.\n b. If the information is useful BUT you still need more details, you MUST search again with <search>...</search>.\n c. If the information is not useful, you MUST search again with <search>...</search>. You MUST NOT provide an answer in this case.\n 5. Provide your final answer in <answer>...</answer> tags. The <answer> tag marks the end of the task. After providing the <answer>, you MUST stop and generate no further text.\n\n Question: where is the setting of charlotte's web?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "barnyard" ] }, "style": "rule" }
{ "index": 476, "need_tools_kwargs": true, "question": "where is the setting of charlotte's web?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "barnyard" ] }, "question": "where is the setting of charlotte's web?" } } } }
null
nq
[ { "content": "You are a helpful and harmless assistant.", "role": "system" }, { "content": "Answer the given question step by step.\n Instructions:\n 1. First, conduct reasoning inside <think> and </think> tags whenever you receive new information.\n 2. If you need external knowledge, you can search using <search>query</search>.\n 3. When you receive search results, evaluate their usefulness and put your judgment inside <judge>Yes</judge> or <judge>No</judge> tags.\n 4. Based on your judgment, follow these strict rules:\n a. If the information is useful AND you now have sufficient information to provide a complete final answer, proceed directly to step 5.\n b. If the information is useful BUT you still need more details, you MUST search again with <search>...</search>.\n c. If the information is not useful, you MUST search again with <search>...</search>. You MUST NOT provide an answer in this case.\n 5. Provide your final answer in <answer>...</answer> tags. The <answer> tag marks the end of the task. After providing the <answer>, you MUST stop and generate no further text.\n\n Question: where does the term elephant in the room come from?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "The New York Times" ] }, "style": "rule" }
{ "index": 477, "need_tools_kwargs": true, "question": "where does the term elephant in the room come from?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "The New York Times" ] }, "question": "where does the term elephant in the room come from?" } } } }
null
nq
[ { "content": "You are a helpful and harmless assistant.", "role": "system" }, { "content": "Answer the given question step by step.\n Instructions:\n 1. First, conduct reasoning inside <think> and </think> tags whenever you receive new information.\n 2. If you need external knowledge, you can search using <search>query</search>.\n 3. When you receive search results, evaluate their usefulness and put your judgment inside <judge>Yes</judge> or <judge>No</judge> tags.\n 4. Based on your judgment, follow these strict rules:\n a. If the information is useful AND you now have sufficient information to provide a complete final answer, proceed directly to step 5.\n b. If the information is useful BUT you still need more details, you MUST search again with <search>...</search>.\n c. If the information is not useful, you MUST search again with <search>...</search>. You MUST NOT provide an answer in this case.\n 5. Provide your final answer in <answer>...</answer> tags. The <answer> tag marks the end of the task. After providing the <answer>, you MUST stop and generate no further text.\n\n Question: when did they start singing god bless america at baseball games?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "September 11 attacks" ] }, "style": "rule" }
{ "index": 478, "need_tools_kwargs": true, "question": "when did they start singing god bless america at baseball games?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "September 11 attacks" ] }, "question": "when did they start singing god bless america at baseball games?" } } } }
null
nq
[ { "content": "You are a helpful and harmless assistant.", "role": "system" }, { "content": "Answer the given question step by step.\n Instructions:\n 1. First, conduct reasoning inside <think> and </think> tags whenever you receive new information.\n 2. If you need external knowledge, you can search using <search>query</search>.\n 3. When you receive search results, evaluate their usefulness and put your judgment inside <judge>Yes</judge> or <judge>No</judge> tags.\n 4. Based on your judgment, follow these strict rules:\n a. If the information is useful AND you now have sufficient information to provide a complete final answer, proceed directly to step 5.\n b. If the information is useful BUT you still need more details, you MUST search again with <search>...</search>.\n c. If the information is not useful, you MUST search again with <search>...</search>. You MUST NOT provide an answer in this case.\n 5. Provide your final answer in <answer>...</answer> tags. The <answer> tag marks the end of the task. After providing the <answer>, you MUST stop and generate no further text.\n\n Question: which country has the greatest loss of life in ww2?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "Soviet Union" ] }, "style": "rule" }
{ "index": 479, "need_tools_kwargs": true, "question": "which country has the greatest loss of life in ww2?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "Soviet Union" ] }, "question": "which country has the greatest loss of life in ww2?" } } } }
null
nq
[ { "content": "You are a helpful and harmless assistant.", "role": "system" }, { "content": "Answer the given question step by step.\n Instructions:\n 1. First, conduct reasoning inside <think> and </think> tags whenever you receive new information.\n 2. If you need external knowledge, you can search using <search>query</search>.\n 3. When you receive search results, evaluate their usefulness and put your judgment inside <judge>Yes</judge> or <judge>No</judge> tags.\n 4. Based on your judgment, follow these strict rules:\n a. If the information is useful AND you now have sufficient information to provide a complete final answer, proceed directly to step 5.\n b. If the information is useful BUT you still need more details, you MUST search again with <search>...</search>.\n c. If the information is not useful, you MUST search again with <search>...</search>. You MUST NOT provide an answer in this case.\n 5. Provide your final answer in <answer>...</answer> tags. The <answer> tag marks the end of the task. After providing the <answer>, you MUST stop and generate no further text.\n\n Question: who played guitar solo on reeling in the years?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "Elliott Randall" ] }, "style": "rule" }
{ "index": 480, "need_tools_kwargs": true, "question": "who played guitar solo on reeling in the years?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "Elliott Randall" ] }, "question": "who played guitar solo on reeling in the years?" } } } }
null
nq
[ { "content": "You are a helpful and harmless assistant.", "role": "system" }, { "content": "Answer the given question step by step.\n Instructions:\n 1. First, conduct reasoning inside <think> and </think> tags whenever you receive new information.\n 2. If you need external knowledge, you can search using <search>query</search>.\n 3. When you receive search results, evaluate their usefulness and put your judgment inside <judge>Yes</judge> or <judge>No</judge> tags.\n 4. Based on your judgment, follow these strict rules:\n a. If the information is useful AND you now have sufficient information to provide a complete final answer, proceed directly to step 5.\n b. If the information is useful BUT you still need more details, you MUST search again with <search>...</search>.\n c. If the information is not useful, you MUST search again with <search>...</search>. You MUST NOT provide an answer in this case.\n 5. Provide your final answer in <answer>...</answer> tags. The <answer> tag marks the end of the task. After providing the <answer>, you MUST stop and generate no further text.\n\n Question: where does the final battle in julius caesar take place?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "in the plain" ] }, "style": "rule" }
{ "index": 481, "need_tools_kwargs": true, "question": "where does the final battle in julius caesar take place?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "in the plain" ] }, "question": "where does the final battle in julius caesar take place?" } } } }
null
nq
[ { "content": "You are a helpful and harmless assistant.", "role": "system" }, { "content": "Answer the given question step by step.\n Instructions:\n 1. First, conduct reasoning inside <think> and </think> tags whenever you receive new information.\n 2. If you need external knowledge, you can search using <search>query</search>.\n 3. When you receive search results, evaluate their usefulness and put your judgment inside <judge>Yes</judge> or <judge>No</judge> tags.\n 4. Based on your judgment, follow these strict rules:\n a. If the information is useful AND you now have sufficient information to provide a complete final answer, proceed directly to step 5.\n b. If the information is useful BUT you still need more details, you MUST search again with <search>...</search>.\n c. If the information is not useful, you MUST search again with <search>...</search>. You MUST NOT provide an answer in this case.\n 5. Provide your final answer in <answer>...</answer> tags. The <answer> tag marks the end of the task. After providing the <answer>, you MUST stop and generate no further text.\n\n Question: who wrote the parks and recreation theme song?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "Gaby Moreno", "Vincent Jones" ] }, "style": "rule" }
{ "index": 482, "need_tools_kwargs": true, "question": "who wrote the parks and recreation theme song?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "Gaby Moreno", "Vincent Jones" ] }, "question": "who wrote the parks and recreation theme song?" } } } }
null
nq
[ { "content": "You are a helpful and harmless assistant.", "role": "system" }, { "content": "Answer the given question step by step.\n Instructions:\n 1. First, conduct reasoning inside <think> and </think> tags whenever you receive new information.\n 2. If you need external knowledge, you can search using <search>query</search>.\n 3. When you receive search results, evaluate their usefulness and put your judgment inside <judge>Yes</judge> or <judge>No</judge> tags.\n 4. Based on your judgment, follow these strict rules:\n a. If the information is useful AND you now have sufficient information to provide a complete final answer, proceed directly to step 5.\n b. If the information is useful BUT you still need more details, you MUST search again with <search>...</search>.\n c. If the information is not useful, you MUST search again with <search>...</search>. You MUST NOT provide an answer in this case.\n 5. Provide your final answer in <answer>...</answer> tags. The <answer> tag marks the end of the task. After providing the <answer>, you MUST stop and generate no further text.\n\n Question: who sang i shall be released the band?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "The Band" ] }, "style": "rule" }
{ "index": 483, "need_tools_kwargs": true, "question": "who sang i shall be released the band?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "The Band" ] }, "question": "who sang i shall be released the band?" } } } }
null
nq
[ { "content": "You are a helpful and harmless assistant.", "role": "system" }, { "content": "Answer the given question step by step.\n Instructions:\n 1. First, conduct reasoning inside <think> and </think> tags whenever you receive new information.\n 2. If you need external knowledge, you can search using <search>query</search>.\n 3. When you receive search results, evaluate their usefulness and put your judgment inside <judge>Yes</judge> or <judge>No</judge> tags.\n 4. Based on your judgment, follow these strict rules:\n a. If the information is useful AND you now have sufficient information to provide a complete final answer, proceed directly to step 5.\n b. If the information is useful BUT you still need more details, you MUST search again with <search>...</search>.\n c. If the information is not useful, you MUST search again with <search>...</search>. You MUST NOT provide an answer in this case.\n 5. Provide your final answer in <answer>...</answer> tags. The <answer> tag marks the end of the task. After providing the <answer>, you MUST stop and generate no further text.\n\n Question: what order do the resident evil films go in?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "Resident Evil (2002)" ] }, "style": "rule" }
{ "index": 484, "need_tools_kwargs": true, "question": "what order do the resident evil films go in?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "Resident Evil (2002)" ] }, "question": "what order do the resident evil films go in?" } } } }
null
nq
[ { "content": "You are a helpful and harmless assistant.", "role": "system" }, { "content": "Answer the given question step by step.\n Instructions:\n 1. First, conduct reasoning inside <think> and </think> tags whenever you receive new information.\n 2. If you need external knowledge, you can search using <search>query</search>.\n 3. When you receive search results, evaluate their usefulness and put your judgment inside <judge>Yes</judge> or <judge>No</judge> tags.\n 4. Based on your judgment, follow these strict rules:\n a. If the information is useful AND you now have sufficient information to provide a complete final answer, proceed directly to step 5.\n b. If the information is useful BUT you still need more details, you MUST search again with <search>...</search>.\n c. If the information is not useful, you MUST search again with <search>...</search>. You MUST NOT provide an answer in this case.\n 5. Provide your final answer in <answer>...</answer> tags. The <answer> tag marks the end of the task. After providing the <answer>, you MUST stop and generate no further text.\n\n Question: who sang i put a spell on you first?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "Jay Hawkins" ] }, "style": "rule" }
{ "index": 485, "need_tools_kwargs": true, "question": "who sang i put a spell on you first?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "Jay Hawkins" ] }, "question": "who sang i put a spell on you first?" } } } }
null
nq
[ { "content": "You are a helpful and harmless assistant.", "role": "system" }, { "content": "Answer the given question step by step.\n Instructions:\n 1. First, conduct reasoning inside <think> and </think> tags whenever you receive new information.\n 2. If you need external knowledge, you can search using <search>query</search>.\n 3. When you receive search results, evaluate their usefulness and put your judgment inside <judge>Yes</judge> or <judge>No</judge> tags.\n 4. Based on your judgment, follow these strict rules:\n a. If the information is useful AND you now have sufficient information to provide a complete final answer, proceed directly to step 5.\n b. If the information is useful BUT you still need more details, you MUST search again with <search>...</search>.\n c. If the information is not useful, you MUST search again with <search>...</search>. You MUST NOT provide an answer in this case.\n 5. Provide your final answer in <answer>...</answer> tags. The <answer> tag marks the end of the task. After providing the <answer>, you MUST stop and generate no further text.\n\n Question: who was given the ten commandments by god?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "Moses" ] }, "style": "rule" }
{ "index": 486, "need_tools_kwargs": true, "question": "who was given the ten commandments by god?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "Moses" ] }, "question": "who was given the ten commandments by god?" } } } }
null
nq
[ { "content": "You are a helpful and harmless assistant.", "role": "system" }, { "content": "Answer the given question step by step.\n Instructions:\n 1. First, conduct reasoning inside <think> and </think> tags whenever you receive new information.\n 2. If you need external knowledge, you can search using <search>query</search>.\n 3. When you receive search results, evaluate their usefulness and put your judgment inside <judge>Yes</judge> or <judge>No</judge> tags.\n 4. Based on your judgment, follow these strict rules:\n a. If the information is useful AND you now have sufficient information to provide a complete final answer, proceed directly to step 5.\n b. If the information is useful BUT you still need more details, you MUST search again with <search>...</search>.\n c. If the information is not useful, you MUST search again with <search>...</search>. You MUST NOT provide an answer in this case.\n 5. Provide your final answer in <answer>...</answer> tags. The <answer> tag marks the end of the task. After providing the <answer>, you MUST stop and generate no further text.\n\n Question: when did ford quit making the 7.3 diesel?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "2003" ] }, "style": "rule" }
{ "index": 487, "need_tools_kwargs": true, "question": "when did ford quit making the 7.3 diesel?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "2003" ] }, "question": "when did ford quit making the 7.3 diesel?" } } } }
null
nq
[ { "content": "You are a helpful and harmless assistant.", "role": "system" }, { "content": "Answer the given question step by step.\n Instructions:\n 1. First, conduct reasoning inside <think> and </think> tags whenever you receive new information.\n 2. If you need external knowledge, you can search using <search>query</search>.\n 3. When you receive search results, evaluate their usefulness and put your judgment inside <judge>Yes</judge> or <judge>No</judge> tags.\n 4. Based on your judgment, follow these strict rules:\n a. If the information is useful AND you now have sufficient information to provide a complete final answer, proceed directly to step 5.\n b. If the information is useful BUT you still need more details, you MUST search again with <search>...</search>.\n c. If the information is not useful, you MUST search again with <search>...</search>. You MUST NOT provide an answer in this case.\n 5. Provide your final answer in <answer>...</answer> tags. The <answer> tag marks the end of the task. After providing the <answer>, you MUST stop and generate no further text.\n\n Question: who plays poussey in orange is the new black?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "Samira Wiley" ] }, "style": "rule" }
{ "index": 488, "need_tools_kwargs": true, "question": "who plays poussey in orange is the new black?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "Samira Wiley" ] }, "question": "who plays poussey in orange is the new black?" } } } }
null
nq
[ { "content": "You are a helpful and harmless assistant.", "role": "system" }, { "content": "Answer the given question step by step.\n Instructions:\n 1. First, conduct reasoning inside <think> and </think> tags whenever you receive new information.\n 2. If you need external knowledge, you can search using <search>query</search>.\n 3. When you receive search results, evaluate their usefulness and put your judgment inside <judge>Yes</judge> or <judge>No</judge> tags.\n 4. Based on your judgment, follow these strict rules:\n a. If the information is useful AND you now have sufficient information to provide a complete final answer, proceed directly to step 5.\n b. If the information is useful BUT you still need more details, you MUST search again with <search>...</search>.\n c. If the information is not useful, you MUST search again with <search>...</search>. You MUST NOT provide an answer in this case.\n 5. Provide your final answer in <answer>...</answer> tags. The <answer> tag marks the end of the task. After providing the <answer>, you MUST stop and generate no further text.\n\n Question: built-in set of machine code instructions of computer are called?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "a computer program" ] }, "style": "rule" }
{ "index": 489, "need_tools_kwargs": true, "question": "built-in set of machine code instructions of computer are called?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "a computer program" ] }, "question": "built-in set of machine code instructions of computer are called?" } } } }
null
nq
[ { "content": "You are a helpful and harmless assistant.", "role": "system" }, { "content": "Answer the given question step by step.\n Instructions:\n 1. First, conduct reasoning inside <think> and </think> tags whenever you receive new information.\n 2. If you need external knowledge, you can search using <search>query</search>.\n 3. When you receive search results, evaluate their usefulness and put your judgment inside <judge>Yes</judge> or <judge>No</judge> tags.\n 4. Based on your judgment, follow these strict rules:\n a. If the information is useful AND you now have sufficient information to provide a complete final answer, proceed directly to step 5.\n b. If the information is useful BUT you still need more details, you MUST search again with <search>...</search>.\n c. If the information is not useful, you MUST search again with <search>...</search>. You MUST NOT provide an answer in this case.\n 5. Provide your final answer in <answer>...</answer> tags. The <answer> tag marks the end of the task. After providing the <answer>, you MUST stop and generate no further text.\n\n Question: what is the latest episode of the walking dead called?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "Wrath" ] }, "style": "rule" }
{ "index": 490, "need_tools_kwargs": true, "question": "what is the latest episode of the walking dead called?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "Wrath" ] }, "question": "what is the latest episode of the walking dead called?" } } } }
null
nq
[ { "content": "You are a helpful and harmless assistant.", "role": "system" }, { "content": "Answer the given question step by step.\n Instructions:\n 1. First, conduct reasoning inside <think> and </think> tags whenever you receive new information.\n 2. If you need external knowledge, you can search using <search>query</search>.\n 3. When you receive search results, evaluate their usefulness and put your judgment inside <judge>Yes</judge> or <judge>No</judge> tags.\n 4. Based on your judgment, follow these strict rules:\n a. If the information is useful AND you now have sufficient information to provide a complete final answer, proceed directly to step 5.\n b. If the information is useful BUT you still need more details, you MUST search again with <search>...</search>.\n c. If the information is not useful, you MUST search again with <search>...</search>. You MUST NOT provide an answer in this case.\n 5. Provide your final answer in <answer>...</answer> tags. The <answer> tag marks the end of the task. After providing the <answer>, you MUST stop and generate no further text.\n\n Question: when did the first episode of steven universe come out?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "November 4, 2013" ] }, "style": "rule" }
{ "index": 491, "need_tools_kwargs": true, "question": "when did the first episode of steven universe come out?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "November 4, 2013" ] }, "question": "when did the first episode of steven universe come out?" } } } }
null
nq
[ { "content": "You are a helpful and harmless assistant.", "role": "system" }, { "content": "Answer the given question step by step.\n Instructions:\n 1. First, conduct reasoning inside <think> and </think> tags whenever you receive new information.\n 2. If you need external knowledge, you can search using <search>query</search>.\n 3. When you receive search results, evaluate their usefulness and put your judgment inside <judge>Yes</judge> or <judge>No</judge> tags.\n 4. Based on your judgment, follow these strict rules:\n a. If the information is useful AND you now have sufficient information to provide a complete final answer, proceed directly to step 5.\n b. If the information is useful BUT you still need more details, you MUST search again with <search>...</search>.\n c. If the information is not useful, you MUST search again with <search>...</search>. You MUST NOT provide an answer in this case.\n 5. Provide your final answer in <answer>...</answer> tags. The <answer> tag marks the end of the task. After providing the <answer>, you MUST stop and generate no further text.\n\n Question: where is tottenham's new ground being built?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "London, England" ] }, "style": "rule" }
{ "index": 492, "need_tools_kwargs": true, "question": "where is tottenham's new ground being built?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "London, England" ] }, "question": "where is tottenham's new ground being built?" } } } }
null
nq
[ { "content": "You are a helpful and harmless assistant.", "role": "system" }, { "content": "Answer the given question step by step.\n Instructions:\n 1. First, conduct reasoning inside <think> and </think> tags whenever you receive new information.\n 2. If you need external knowledge, you can search using <search>query</search>.\n 3. When you receive search results, evaluate their usefulness and put your judgment inside <judge>Yes</judge> or <judge>No</judge> tags.\n 4. Based on your judgment, follow these strict rules:\n a. If the information is useful AND you now have sufficient information to provide a complete final answer, proceed directly to step 5.\n b. If the information is useful BUT you still need more details, you MUST search again with <search>...</search>.\n c. If the information is not useful, you MUST search again with <search>...</search>. You MUST NOT provide an answer in this case.\n 5. Provide your final answer in <answer>...</answer> tags. The <answer> tag marks the end of the task. After providing the <answer>, you MUST stop and generate no further text.\n\n Question: where does the ferry across the mersey go to?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "Birkenhead/Wallasey", "Liverpool" ] }, "style": "rule" }
{ "index": 493, "need_tools_kwargs": true, "question": "where does the ferry across the mersey go to?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "Birkenhead/Wallasey", "Liverpool" ] }, "question": "where does the ferry across the mersey go to?" } } } }
null
nq
[ { "content": "You are a helpful and harmless assistant.", "role": "system" }, { "content": "Answer the given question step by step.\n Instructions:\n 1. First, conduct reasoning inside <think> and </think> tags whenever you receive new information.\n 2. If you need external knowledge, you can search using <search>query</search>.\n 3. When you receive search results, evaluate their usefulness and put your judgment inside <judge>Yes</judge> or <judge>No</judge> tags.\n 4. Based on your judgment, follow these strict rules:\n a. If the information is useful AND you now have sufficient information to provide a complete final answer, proceed directly to step 5.\n b. If the information is useful BUT you still need more details, you MUST search again with <search>...</search>.\n c. If the information is not useful, you MUST search again with <search>...</search>. You MUST NOT provide an answer in this case.\n 5. Provide your final answer in <answer>...</answer> tags. The <answer> tag marks the end of the task. After providing the <answer>, you MUST stop and generate no further text.\n\n Question: who is the actor that plays stefan salvatore?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "Paul Wesley" ] }, "style": "rule" }
{ "index": 494, "need_tools_kwargs": true, "question": "who is the actor that plays stefan salvatore?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "Paul Wesley" ] }, "question": "who is the actor that plays stefan salvatore?" } } } }
null
nq
[ { "content": "You are a helpful and harmless assistant.", "role": "system" }, { "content": "Answer the given question step by step.\n Instructions:\n 1. First, conduct reasoning inside <think> and </think> tags whenever you receive new information.\n 2. If you need external knowledge, you can search using <search>query</search>.\n 3. When you receive search results, evaluate their usefulness and put your judgment inside <judge>Yes</judge> or <judge>No</judge> tags.\n 4. Based on your judgment, follow these strict rules:\n a. If the information is useful AND you now have sufficient information to provide a complete final answer, proceed directly to step 5.\n b. If the information is useful BUT you still need more details, you MUST search again with <search>...</search>.\n c. If the information is not useful, you MUST search again with <search>...</search>. You MUST NOT provide an answer in this case.\n 5. Provide your final answer in <answer>...</answer> tags. The <answer> tag marks the end of the task. After providing the <answer>, you MUST stop and generate no further text.\n\n Question: where did they film the movie cabin fever?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "Portland, Oregon" ] }, "style": "rule" }
{ "index": 495, "need_tools_kwargs": true, "question": "where did they film the movie cabin fever?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "Portland, Oregon" ] }, "question": "where did they film the movie cabin fever?" } } } }
null
nq
[ { "content": "You are a helpful and harmless assistant.", "role": "system" }, { "content": "Answer the given question step by step.\n Instructions:\n 1. First, conduct reasoning inside <think> and </think> tags whenever you receive new information.\n 2. If you need external knowledge, you can search using <search>query</search>.\n 3. When you receive search results, evaluate their usefulness and put your judgment inside <judge>Yes</judge> or <judge>No</judge> tags.\n 4. Based on your judgment, follow these strict rules:\n a. If the information is useful AND you now have sufficient information to provide a complete final answer, proceed directly to step 5.\n b. If the information is useful BUT you still need more details, you MUST search again with <search>...</search>.\n c. If the information is not useful, you MUST search again with <search>...</search>. You MUST NOT provide an answer in this case.\n 5. Provide your final answer in <answer>...</answer> tags. The <answer> tag marks the end of the task. After providing the <answer>, you MUST stop and generate no further text.\n\n Question: when did season 2 of the last kingdom come out?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "16 March 2017" ] }, "style": "rule" }
{ "index": 496, "need_tools_kwargs": true, "question": "when did season 2 of the last kingdom come out?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "16 March 2017" ] }, "question": "when did season 2 of the last kingdom come out?" } } } }
null
nq
[ { "content": "You are a helpful and harmless assistant.", "role": "system" }, { "content": "Answer the given question step by step.\n Instructions:\n 1. First, conduct reasoning inside <think> and </think> tags whenever you receive new information.\n 2. If you need external knowledge, you can search using <search>query</search>.\n 3. When you receive search results, evaluate their usefulness and put your judgment inside <judge>Yes</judge> or <judge>No</judge> tags.\n 4. Based on your judgment, follow these strict rules:\n a. If the information is useful AND you now have sufficient information to provide a complete final answer, proceed directly to step 5.\n b. If the information is useful BUT you still need more details, you MUST search again with <search>...</search>.\n c. If the information is not useful, you MUST search again with <search>...</search>. You MUST NOT provide an answer in this case.\n 5. Provide your final answer in <answer>...</answer> tags. The <answer> tag marks the end of the task. After providing the <answer>, you MUST stop and generate no further text.\n\n Question: when did australia sign the declaration of human rights?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "10 December 1948" ] }, "style": "rule" }
{ "index": 497, "need_tools_kwargs": true, "question": "when did australia sign the declaration of human rights?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "10 December 1948" ] }, "question": "when did australia sign the declaration of human rights?" } } } }
null
nq
[ { "content": "You are a helpful and harmless assistant.", "role": "system" }, { "content": "Answer the given question step by step.\n Instructions:\n 1. First, conduct reasoning inside <think> and </think> tags whenever you receive new information.\n 2. If you need external knowledge, you can search using <search>query</search>.\n 3. When you receive search results, evaluate their usefulness and put your judgment inside <judge>Yes</judge> or <judge>No</judge> tags.\n 4. Based on your judgment, follow these strict rules:\n a. If the information is useful AND you now have sufficient information to provide a complete final answer, proceed directly to step 5.\n b. If the information is useful BUT you still need more details, you MUST search again with <search>...</search>.\n c. If the information is not useful, you MUST search again with <search>...</search>. You MUST NOT provide an answer in this case.\n 5. Provide your final answer in <answer>...</answer> tags. The <answer> tag marks the end of the task. After providing the <answer>, you MUST stop and generate no further text.\n\n Question: name the group of seven stars seen in the sky?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "Pleiades" ] }, "style": "rule" }
{ "index": 498, "need_tools_kwargs": true, "question": "name the group of seven stars seen in the sky?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "Pleiades" ] }, "question": "name the group of seven stars seen in the sky?" } } } }
null
nq
[ { "content": "You are a helpful and harmless assistant.", "role": "system" }, { "content": "Answer the given question step by step.\n Instructions:\n 1. First, conduct reasoning inside <think> and </think> tags whenever you receive new information.\n 2. If you need external knowledge, you can search using <search>query</search>.\n 3. When you receive search results, evaluate their usefulness and put your judgment inside <judge>Yes</judge> or <judge>No</judge> tags.\n 4. Based on your judgment, follow these strict rules:\n a. If the information is useful AND you now have sufficient information to provide a complete final answer, proceed directly to step 5.\n b. If the information is useful BUT you still need more details, you MUST search again with <search>...</search>.\n c. If the information is not useful, you MUST search again with <search>...</search>. You MUST NOT provide an answer in this case.\n 5. Provide your final answer in <answer>...</answer> tags. The <answer> tag marks the end of the task. After providing the <answer>, you MUST stop and generate no further text.\n\n Question: the ruler of the holy roman empire that split his empire in half between his two sons was who?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "Theodosius I" ] }, "style": "rule" }
{ "index": 499, "need_tools_kwargs": true, "question": "the ruler of the holy roman empire that split his empire in half between his two sons was who?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "Theodosius I" ] }, "question": "the ruler of the holy roman empire that split his empire in half between his two sons was who?" } } } }
null