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: who played david in the assassination of versace?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "Terry Sweeney" ] }, "style": "rule" }
{ "index": 500, "need_tools_kwargs": true, "question": "who played david in the assassination of versace?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "Terry Sweeney" ] }, "question": "who played david in the assassination of versace?" } } } }
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 in the super bowl this year?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "New England Patriots", "Atlanta Falcons" ] }, "style": "rule" }
{ "index": 501, "need_tools_kwargs": true, "question": "who played in the super bowl this year?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "New England Patriots", "Atlanta Falcons" ] }, "question": "who played in the super bowl this year?" } } } }
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 california become part of the us?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "on September 9, 1850" ] }, "style": "rule" }
{ "index": 502, "need_tools_kwargs": true, "question": "when did california become part of the us?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "on September 9, 1850" ] }, "question": "when did california become part of the us?" } } } }
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 most olympic medals in history?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "Michael Phelps" ] }, "style": "rule" }
{ "index": 503, "need_tools_kwargs": true, "question": "who won the most olympic medals in history?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "Michael Phelps" ] }, "question": "who won the most olympic medals in history?" } } } }
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 were the main proponents of amending the constitution to include a bill of rights?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "James Madison", "Alexander Hamilton" ] }, "style": "rule" }
{ "index": 504, "need_tools_kwargs": true, "question": "who were the main proponents of amending the constitution to include a bill of rights?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "James Madison", "Alexander Hamilton" ] }, "question": "who were the main proponents of amending the constitution to include a bill of 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: who has the fastest roller coaster in the world?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "Ferrari World Abu Dhabi" ] }, "style": "rule" }
{ "index": 505, "need_tools_kwargs": true, "question": "who has the fastest roller coaster in the world?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "Ferrari World Abu Dhabi" ] }, "question": "who has the fastest roller coaster 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: what is the meaning of ragi in punjabi?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "(finger millet) flour" ] }, "style": "rule" }
{ "index": 506, "need_tools_kwargs": true, "question": "what is the meaning of ragi in punjabi?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "(finger millet) flour" ] }, "question": "what is the meaning of ragi in punjabi?" } } } }
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 early textile mills have to be built?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "rural locations" ] }, "style": "rule" }
{ "index": 507, "need_tools_kwargs": true, "question": "where did early textile mills have to be built?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "rural locations" ] }, "question": "where did early textile mills have to be 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: when was the last time a foreign leader addressed congress?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "June 8, 2016" ] }, "style": "rule" }
{ "index": 508, "need_tools_kwargs": true, "question": "when was the last time a foreign leader addressed congress?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "June 8, 2016" ] }, "question": "when was the last time a foreign leader addressed congress?" } } } }
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 heat come from in our body?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "homeostasis" ] }, "style": "rule" }
{ "index": 509, "need_tools_kwargs": true, "question": "where does heat come from in our body?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "homeostasis" ] }, "question": "where does heat come from in our body?" } } } }
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 jem and the holograms come out?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "1985" ] }, "style": "rule" }
{ "index": 510, "need_tools_kwargs": true, "question": "when did jem and the holograms come out?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "1985" ] }, "question": "when did jem and the holograms 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: who was in charge of the lost colony?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "Sir Walter Raleigh" ] }, "style": "rule" }
{ "index": 511, "need_tools_kwargs": true, "question": "who was in charge of the lost colony?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "Sir Walter Raleigh" ] }, "question": "who was in charge of the lost colony?" } } } }
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 person to go skydiving?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "André-Jacques Garnerin" ] }, "style": "rule" }
{ "index": 512, "need_tools_kwargs": true, "question": "who was the first person to go skydiving?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "André-Jacques Garnerin" ] }, "question": "who was the first person to go skydiving?" } } } }
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 sun baby in the original run of teletubbies?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "Jess Smith" ] }, "style": "rule" }
{ "index": 513, "need_tools_kwargs": true, "question": "who played the sun baby in the original run of teletubbies?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "Jess Smith" ] }, "question": "who played the sun baby in the original run of teletubbies?" } } } }
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 amy grant and i can only imagine?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "Nicole DuPort" ] }, "style": "rule" }
{ "index": 514, "need_tools_kwargs": true, "question": "who played amy grant and i can only imagine?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "Nicole DuPort" ] }, "question": "who played amy grant and i can only imagine?" } } } }
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: 4. what does c stand for in case of disk scheduling algorithms?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "Circular Elevator Algorithm" ] }, "style": "rule" }
{ "index": 515, "need_tools_kwargs": true, "question": "4. what does c stand for in case of disk scheduling algorithms?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "Circular Elevator Algorithm" ] }, "question": "4. what does c stand for in case of disk scheduling algorithms?" } } } }
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 second book of the old testament?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "Exodus" ] }, "style": "rule" }
{ "index": 516, "need_tools_kwargs": true, "question": "what is the second book of the old testament?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "Exodus" ] }, "question": "what is the second book of the old testament?" } } } }
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: badrinath ki dulhania alia bhatt name in movie?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "Vaidehi Trivedi" ] }, "style": "rule" }
{ "index": 517, "need_tools_kwargs": true, "question": "badrinath ki dulhania alia bhatt name in movie?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "Vaidehi Trivedi" ] }, "question": "badrinath ki dulhania alia bhatt name in movie?" } } } }
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 cbs channel is the late late show on?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "CBS" ] }, "style": "rule" }
{ "index": 518, "need_tools_kwargs": true, "question": "what cbs channel is the late late show on?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "CBS" ] }, "question": "what cbs channel is the late late show on?" } } } }
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 contract in the nba?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "Stephen Curry" ] }, "style": "rule" }
{ "index": 519, "need_tools_kwargs": true, "question": "who has the largest contract in the nba?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "Stephen Curry" ] }, "question": "who has the largest contract in the nba?" } } } }
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 notes in when you reach me?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "Rebecca Stead" ] }, "style": "rule" }
{ "index": 520, "need_tools_kwargs": true, "question": "who wrote the notes in when you reach me?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "Rebecca Stead" ] }, "question": "who wrote the notes in when you reach me?" } } } }
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 law or bill that changes the legal consequences of acts committed on the legal status of facts?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "ex post facto law" ] }, "style": "rule" }
{ "index": 521, "need_tools_kwargs": true, "question": "the law or bill that changes the legal consequences of acts committed on the legal status of facts?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "ex post facto law" ] }, "question": "the law or bill that changes the legal consequences of acts committed on the legal status of facts?" } } } }
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 emily prentiss return to criminal minds?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "August 30, 2016" ] }, "style": "rule" }
{ "index": 522, "need_tools_kwargs": true, "question": "when did emily prentiss return to criminal minds?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "August 30, 2016" ] }, "question": "when did emily prentiss return to criminal minds?" } } } }
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 place beyond the pines set?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "1997" ] }, "style": "rule" }
{ "index": 523, "need_tools_kwargs": true, "question": "when is the place beyond the pines set?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "1997" ] }, "question": "when is the place beyond the pines set?" } } } }
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 national science day is observed every year on 28 february in honour of?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "Sir Chandrashekhara Venkata Raman" ] }, "style": "rule" }
{ "index": 524, "need_tools_kwargs": true, "question": "the national science day is observed every year on 28 february in honour of?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "Sir Chandrashekhara Venkata Raman" ] }, "question": "the national science day is observed every year on 28 february in honour 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 played in the ghost and mrs. muir?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "Kellie Flanagan", "Harlen Carraher", "Charles Nelson Reilly", "Reta Shaw", "Edward Mulhare", "Hope Lange" ] }, "style": "rule" }
{ "index": 525, "need_tools_kwargs": true, "question": "who played in the ghost and mrs. muir?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "Kellie Flanagan", "Harlen Carraher", "Charles Nelson Reilly", "Reta Shaw", "Edward Mulhare", "Hope Lange" ] }, "question": "who played in the ghost and mrs. muir?" } } } }
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 bartender in how i met your mother?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "Joe Nieves" ] }, "style": "rule" }
{ "index": 526, "need_tools_kwargs": true, "question": "who plays the bartender in how i met your mother?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "Joe Nieves" ] }, "question": "who plays the bartender in how i met your mother?" } } } }
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 midwest express go out of business?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "in late 2011" ] }, "style": "rule" }
{ "index": 527, "need_tools_kwargs": true, "question": "when did midwest express go out of business?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "in late 2011" ] }, "question": "when did midwest express 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: h r pufnstuf what does hr stand for?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "\"Royal Highness\" backwards" ] }, "style": "rule" }
{ "index": 528, "need_tools_kwargs": true, "question": "h r pufnstuf what does hr stand for?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "\"Royal Highness\" backwards" ] }, "question": "h r pufnstuf what does hr stand for?" } } } }
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: drugs that are substrates of cytochrome p450 3a4 and cyp2d6?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "ciclosporin (cyclosporin)", "acetaminophen", "diazepam", "codeine", "erythromycin" ] }, "style": "rule" }
{ "index": 529, "need_tools_kwargs": true, "question": "drugs that are substrates of cytochrome p450 3a4 and cyp2d6?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "ciclosporin (cyclosporin)", "acetaminophen", "diazepam", "codeine", "erythromycin" ] }, "question": "drugs that are substrates of cytochrome p450 3a4 and cyp2d6?" } } } }
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 founder of portuguese power in india?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "Francisco de Almeida" ] }, "style": "rule" }
{ "index": 530, "need_tools_kwargs": true, "question": "who is the founder of portuguese power in india?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "Francisco de Almeida" ] }, "question": "who is the founder of portuguese power in india?" } } } }
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 the theme song for as told by ginger?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "Macy Gray" ] }, "style": "rule" }
{ "index": 531, "need_tools_kwargs": true, "question": "who sang the theme song for as told by ginger?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "Macy Gray" ] }, "question": "who sang the theme song for as told by ginger?" } } } }
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 far is paradise ca from chico ca?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "10 miles" ] }, "style": "rule" }
{ "index": 532, "need_tools_kwargs": true, "question": "how far is paradise ca from chico ca?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "10 miles" ] }, "question": "how far is paradise ca from chico ca?" } } } }
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 screenplay for three billboards outside ebbing missouri?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "Martin McDonagh" ] }, "style": "rule" }
{ "index": 533, "need_tools_kwargs": true, "question": "who wrote the screenplay for three billboards outside ebbing missouri?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "Martin McDonagh" ] }, "question": "who wrote the screenplay for three billboards outside ebbing missouri?" } } } }
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 does senior airman mean in the air force?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "fourth enlisted rank" ] }, "style": "rule" }
{ "index": 534, "need_tools_kwargs": true, "question": "what does senior airman mean in the air force?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "fourth enlisted rank" ] }, "question": "what does senior airman mean in the air force?" } } } }
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 became the governer of bengal in 1758?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "Robert Clive" ] }, "style": "rule" }
{ "index": 535, "need_tools_kwargs": true, "question": "who became the governer of bengal in 1758?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "Robert Clive" ] }, "question": "who became the governer of bengal in 1758?" } } } }
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 the theme song to mystery science theater 3000?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "Hodgson" ] }, "style": "rule" }
{ "index": 536, "need_tools_kwargs": true, "question": "who sings the theme song to mystery science theater 3000?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "Hodgson" ] }, "question": "who sings the theme song to mystery science theater 3000?" } } } }
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 gowen on when calls the heart?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "Martin Cummins" ] }, "style": "rule" }
{ "index": 537, "need_tools_kwargs": true, "question": "who plays gowen on when calls the heart?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "Martin Cummins" ] }, "question": "who plays gowen on when calls the heart?" } } } }
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's the last time the vikings were in the playoffs?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "2015" ] }, "style": "rule" }
{ "index": 538, "need_tools_kwargs": true, "question": "when's the last time the vikings were in the playoffs?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "2015" ] }, "question": "when's the last time the vikings were in the playoffs?" } } } }
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 song the air that i breathe?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "Mike Hazlewood", "Albert Hammond" ] }, "style": "rule" }
{ "index": 539, "need_tools_kwargs": true, "question": "who wrote the song the air that i breathe?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "Mike Hazlewood", "Albert Hammond" ] }, "question": "who wrote the song the air that i breathe?" } } } }
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 mrs. potts in the new beauty and the beast movie?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "Emma Thompson" ] }, "style": "rule" }
{ "index": 540, "need_tools_kwargs": true, "question": "who plays mrs. potts in the new beauty and the beast movie?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "Emma Thompson" ] }, "question": "who plays mrs. potts in the new beauty and the beast movie?" } } } }
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 song i've been everywhere man?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "Australian country singer Geoff Mack" ] }, "style": "rule" }
{ "index": 541, "need_tools_kwargs": true, "question": "who wrote the song i've been everywhere man?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "Australian country singer Geoff Mack" ] }, "question": "who wrote the song i've been everywhere man?" } } } }
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 author of land of stories?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "Chris Colfer" ] }, "style": "rule" }
{ "index": 542, "need_tools_kwargs": true, "question": "who is the author of land of stories?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "Chris Colfer" ] }, "question": "who is the author of land of stories?" } } } }
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 title fuehrer was applied to which famous dictator?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "Adolf Hitler" ] }, "style": "rule" }
{ "index": 543, "need_tools_kwargs": true, "question": "the title fuehrer was applied to which famous dictator?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "Adolf Hitler" ] }, "question": "the title fuehrer was applied to which famous dictator?" } } } }
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 last ncaa football game come out?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "July 9, 2013" ] }, "style": "rule" }
{ "index": 544, "need_tools_kwargs": true, "question": "when did the last ncaa football game come out?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "July 9, 2013" ] }, "question": "when did the last ncaa football game 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: who were judges on the voice last season?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "Jennifer Hudson", "Miley Cyrus", "Blake Shelton", "Adam Levine" ] }, "style": "rule" }
{ "index": 545, "need_tools_kwargs": true, "question": "who were judges on the voice last season?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "Jennifer Hudson", "Miley Cyrus", "Blake Shelton", "Adam Levine" ] }, "question": "who were judges on the voice last season?" } } } }
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 hours for a full time job?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "35 to 40" ] }, "style": "rule" }
{ "index": 546, "need_tools_kwargs": true, "question": "how many hours for a full time job?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "35 to 40" ] }, "question": "how many hours for a full time job?" } } } }
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's the latest version of google chrome?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "63.0" ] }, "style": "rule" }
{ "index": 547, "need_tools_kwargs": true, "question": "what's the latest version of google chrome?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "63.0" ] }, "question": "what's the latest version of google chrome?" } } } }
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 chargers make the super bowl?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "Super Bowl XXIX" ] }, "style": "rule" }
{ "index": 548, "need_tools_kwargs": true, "question": "when did the chargers make the super bowl?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "Super Bowl XXIX" ] }, "question": "when did the chargers make the super bowl?" } } } }
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 girl in blackbear do re mi?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "Gucci Mane" ] }, "style": "rule" }
{ "index": 549, "need_tools_kwargs": true, "question": "who is the girl in blackbear do re mi?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "Gucci Mane" ] }, "question": "who is the girl in blackbear do re mi?" } } } }
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 said better to reign in hell than serve in heaven?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "Satan" ] }, "style": "rule" }
{ "index": 550, "need_tools_kwargs": true, "question": "who said better to reign in hell than serve in heaven?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "Satan" ] }, "question": "who said better to reign in hell than serve in heaven?" } } } }
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 greys anatomy season 14 be released?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "September 28, 2017" ] }, "style": "rule" }
{ "index": 551, "need_tools_kwargs": true, "question": "when will greys anatomy season 14 be released?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "September 28, 2017" ] }, "question": "when will greys anatomy season 14 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 wins at the end of grudge match?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "Razor" ] }, "style": "rule" }
{ "index": 552, "need_tools_kwargs": true, "question": "who wins at the end of grudge match?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "Razor" ] }, "question": "who wins at the end of grudge match?" } } } }
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 narrated the war of the worlds radio broadcast?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "Orson Welles" ] }, "style": "rule" }
{ "index": 553, "need_tools_kwargs": true, "question": "who narrated the war of the worlds radio broadcast?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "Orson Welles" ] }, "question": "who narrated the war of the worlds radio broadcast?" } } } }
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 been chosen as the best actor in the 65th national flim awards 2018?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "Riddhi Sen" ] }, "style": "rule" }
{ "index": 554, "need_tools_kwargs": true, "question": "who has been chosen as the best actor in the 65th national flim awards 2018?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "Riddhi Sen" ] }, "question": "who has been chosen as the best actor in the 65th national flim awards 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: where does the last name waters originate from?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "Wales and Yorkshire" ] }, "style": "rule" }
{ "index": 555, "need_tools_kwargs": true, "question": "where does the last name waters originate from?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "Wales and Yorkshire" ] }, "question": "where does the last name waters originate 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: who are the living presidents of the united states?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "Donald Trump", "Bill Clinton", "Jimmy Carter", "George W. Bush", "Barack Obama", "George H. W. Bush" ] }, "style": "rule" }
{ "index": 556, "need_tools_kwargs": true, "question": "who are the living presidents of the united states?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "Donald Trump", "Bill Clinton", "Jimmy Carter", "George W. Bush", "Barack Obama", "George H. W. Bush" ] }, "question": "who are the living presidents of the united states?" } } } }
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 was the original tomb raider game released?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "1996" ] }, "style": "rule" }
{ "index": 557, "need_tools_kwargs": true, "question": "when was the original tomb raider game released?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "1996" ] }, "question": "when was the original tomb raider game 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 first person to win strictly come dancing?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "Natasha Kaplinsky", "Brendan Cole" ] }, "style": "rule" }
{ "index": 558, "need_tools_kwargs": true, "question": "who was the first person to win strictly come dancing?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "Natasha Kaplinsky", "Brendan Cole" ] }, "question": "who was the first person to win strictly come dancing?" } } } }
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: as a result of ach stimulation calcium ions are released from the?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "the extracellular fluid" ] }, "style": "rule" }
{ "index": 559, "need_tools_kwargs": true, "question": "as a result of ach stimulation calcium ions are released from the?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "the extracellular fluid" ] }, "question": "as a result of ach stimulation calcium ions are released from the?" } } } }
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 was under god added to pledge of allegence?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "in 1954" ] }, "style": "rule" }
{ "index": 560, "need_tools_kwargs": true, "question": "when was under god added to pledge of allegence?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "in 1954" ] }, "question": "when was under god added to pledge of allegence?" } } } }
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 big funnel web spider located?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "Jamberoo, New South Wales" ] }, "style": "rule" }
{ "index": 561, "need_tools_kwargs": true, "question": "where is the big funnel web spider located?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "Jamberoo, New South Wales" ] }, "question": "where is the big funnel web spider located?" } } } }
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 zuzu in it's a wonderful life?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "Karolyn Grimes" ] }, "style": "rule" }
{ "index": 562, "need_tools_kwargs": true, "question": "who played zuzu in it's a wonderful life?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "Karolyn Grimes" ] }, "question": "who played zuzu in it's a wonderful life?" } } } }
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 punter for the dallas cowboys?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "Chris Jones" ] }, "style": "rule" }
{ "index": 563, "need_tools_kwargs": true, "question": "who is the punter for the dallas cowboys?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "Chris Jones" ] }, "question": "who is the punter for the dallas cowboys?" } } } }
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: thespian' is used to describe this ancient greek artist who is considered to be the first actor?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "Thespis" ] }, "style": "rule" }
{ "index": 564, "need_tools_kwargs": true, "question": "thespian' is used to describe this ancient greek artist who is considered to be the first actor?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "Thespis" ] }, "question": "thespian' is used to describe this ancient greek artist who is considered to be the first actor?" } } } }
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 killing of a sacred deer filmed?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "Cincinnati" ] }, "style": "rule" }
{ "index": 565, "need_tools_kwargs": true, "question": "where is killing of a sacred deer filmed?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "Cincinnati" ] }, "question": "where is killing of a sacred deer filmed?" } } } }
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 change the definition of gender?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "1955" ] }, "style": "rule" }
{ "index": 566, "need_tools_kwargs": true, "question": "when did they change the definition of gender?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "1955" ] }, "question": "when did they change the definition of gender?" } } } }
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 attorney general for new jersey?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "Gurbir Grewal" ] }, "style": "rule" }
{ "index": 567, "need_tools_kwargs": true, "question": "who is the attorney general for new jersey?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "Gurbir Grewal" ] }, "question": "who is the attorney general for new jersey?" } } } }
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 most home runs all time?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "Barry Bonds" ] }, "style": "rule" }
{ "index": 568, "need_tools_kwargs": true, "question": "who has the most home runs all time?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "Barry Bonds" ] }, "question": "who has the most home runs 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: when is the new episode of pll coming out?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "April 2017" ] }, "style": "rule" }
{ "index": 569, "need_tools_kwargs": true, "question": "when is the new episode of pll coming out?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "April 2017" ] }, "question": "when is the new episode of pll 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: how many judges are there in the book of judges in the bible?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "twelve" ] }, "style": "rule" }
{ "index": 570, "need_tools_kwargs": true, "question": "how many judges are there in the book of judges in the bible?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "twelve" ] }, "question": "how many judges are there in the book of judges in the bible?" } } } }
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 worked as a constitutional advisor for constitutional assembly?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "B. N. Rau" ] }, "style": "rule" }
{ "index": 571, "need_tools_kwargs": true, "question": "who worked as a constitutional advisor for constitutional assembly?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "B. N. Rau" ] }, "question": "who worked as a constitutional advisor for constitutional assembly?" } } } }
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 song peanut butter jelly time?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "Buckwheat Boyz" ] }, "style": "rule" }
{ "index": 572, "need_tools_kwargs": true, "question": "who wrote the song peanut butter jelly time?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "Buckwheat Boyz" ] }, "question": "who wrote the song peanut butter jelly 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 did usc lose to this year in football?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "Notre Dame" ] }, "style": "rule" }
{ "index": 573, "need_tools_kwargs": true, "question": "who did usc lose to this year in football?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "Notre Dame" ] }, "question": "who did usc lose to this year in football?" } } } }
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 judicial branch of the eu located?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "Luxembourg" ] }, "style": "rule" }
{ "index": 574, "need_tools_kwargs": true, "question": "where is the judicial branch of the eu located?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "Luxembourg" ] }, "question": "where is the judicial branch of the eu located?" } } } }
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 president of dominican republic live?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "The National Palace" ] }, "style": "rule" }
{ "index": 575, "need_tools_kwargs": true, "question": "where does the president of dominican republic live?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "The National Palace" ] }, "question": "where does the president of dominican republic live?" } } } }
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 try to love one another right now?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "The Youngbloods" ] }, "style": "rule" }
{ "index": 576, "need_tools_kwargs": true, "question": "who sings try to love one another right now?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "The Youngbloods" ] }, "question": "who sings try to love one another right 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: how far is walterboro sc from summerville sc?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "32 miles" ] }, "style": "rule" }
{ "index": 577, "need_tools_kwargs": true, "question": "how far is walterboro sc from summerville sc?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "32 miles" ] }, "question": "how far is walterboro sc from summerville sc?" } } } }
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 diary of a wimpy kid movie come out?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "2010" ] }, "style": "rule" }
{ "index": 578, "need_tools_kwargs": true, "question": "when did the first diary of a wimpy kid movie come out?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "2010" ] }, "question": "when did the first diary of a wimpy kid movie 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: who is the he referred to so often in part three of the declaration?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "King George III" ] }, "style": "rule" }
{ "index": 579, "need_tools_kwargs": true, "question": "who is the he referred to so often in part three of the declaration?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "King George III" ] }, "question": "who is the he referred to so often in part three of the declaration?" } } } }
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 declared war on japan after pearl harbor?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "the United States Congress" ] }, "style": "rule" }
{ "index": 580, "need_tools_kwargs": true, "question": "who declared war on japan after pearl harbor?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "the United States Congress" ] }, "question": "who declared war on japan after pearl harbor?" } } } }
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 much does it cost to renew a cuban passport?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "US$500" ] }, "style": "rule" }
{ "index": 581, "need_tools_kwargs": true, "question": "how much does it cost to renew a cuban passport?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "US$500" ] }, "question": "how much does it cost to renew a cuban passport?" } } } }
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 father in rumor has it?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "Richard Jenkins" ] }, "style": "rule" }
{ "index": 582, "need_tools_kwargs": true, "question": "who is the father in rumor has it?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "Richard Jenkins" ] }, "question": "who is the father in rumor has it?" } } } }
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's at the luxor in las vegas?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "Criss Angel Mindfreak Live", "Fantasy", "Carrot Top", "Blue Man Group" ] }, "style": "rule" }
{ "index": 583, "need_tools_kwargs": true, "question": "what's at the luxor in las vegas?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "Criss Angel Mindfreak Live", "Fantasy", "Carrot Top", "Blue Man Group" ] }, "question": "what's at the luxor in las vegas?" } } } }
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 oldest wrestler still wrestling in wwe?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "Dory Funk Jr." ] }, "style": "rule" }
{ "index": 584, "need_tools_kwargs": true, "question": "who is the oldest wrestler still wrestling in wwe?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "Dory Funk Jr." ] }, "question": "who is the oldest wrestler still wrestling in wwe?" } } } }
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 fort leonard wood mo on a map?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "the Missouri Ozarks" ] }, "style": "rule" }
{ "index": 585, "need_tools_kwargs": true, "question": "where is fort leonard wood mo on a map?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "the Missouri Ozarks" ] }, "question": "where is fort leonard wood mo on a map?" } } } }
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 new technology helped win the war in the pacific?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "the atomic bomb" ] }, "style": "rule" }
{ "index": 586, "need_tools_kwargs": true, "question": "what new technology helped win the war in the pacific?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "the atomic bomb" ] }, "question": "what new technology helped win the war in the pacific?" } } } }
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 fa premier league is owned by 20 shareholders who are?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "the 20 member clubs" ] }, "style": "rule" }
{ "index": 587, "need_tools_kwargs": true, "question": "the fa premier league is owned by 20 shareholders who are?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "the 20 member clubs" ] }, "question": "the fa premier league is owned by 20 shareholders who are?" } } } }
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 the first fleet set up camp?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "Botany Bay" ] }, "style": "rule" }
{ "index": 588, "need_tools_kwargs": true, "question": "where did the first fleet set up camp?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "Botany Bay" ] }, "question": "where did the first fleet set up camp?" } } } }
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: all things being equal... is an expression from which school of thought?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "the logical empiricist" ] }, "style": "rule" }
{ "index": 589, "need_tools_kwargs": true, "question": "all things being equal... is an expression from which school of thought?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "the logical empiricist" ] }, "question": "all things being equal... is an expression from which school of thought?" } } } }
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 was india capital shifted from calcutta to delhi?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "12 December 1911" ] }, "style": "rule" }
{ "index": 590, "need_tools_kwargs": true, "question": "when was india capital shifted from calcutta to delhi?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "12 December 1911" ] }, "question": "when was india capital shifted from calcutta to delhi?" } } } }
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 invaded roman territory in the third century?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "Goths", "the Alamanni" ] }, "style": "rule" }
{ "index": 591, "need_tools_kwargs": true, "question": "who invaded roman territory in the third century?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "Goths", "the Alamanni" ] }, "question": "who invaded roman territory in the third century?" } } } }
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 lumiere on beauty and the beast?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "Ewan McGregor" ] }, "style": "rule" }
{ "index": 592, "need_tools_kwargs": true, "question": "who plays lumiere on beauty and the beast?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "Ewan McGregor" ] }, "question": "who plays lumiere on beauty and the beast?" } } } }
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 cindy lou who in the grinch stole christmas?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "Taylor Michel Momsen" ] }, "style": "rule" }
{ "index": 593, "need_tools_kwargs": true, "question": "who played cindy lou who in the grinch stole christmas?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "Taylor Michel Momsen" ] }, "question": "who played cindy lou who in 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: how many times did dale earnhardt win daytona?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "76" ] }, "style": "rule" }
{ "index": 594, "need_tools_kwargs": true, "question": "how many times did dale earnhardt win daytona?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "76" ] }, "question": "how many times did dale earnhardt win daytona?" } } } }
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 race is aaron judge of the new york yankees?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "African-American" ] }, "style": "rule" }
{ "index": 595, "need_tools_kwargs": true, "question": "what race is aaron judge of the new york yankees?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "African-American" ] }, "question": "what race is aaron judge of the new york yankees?" } } } }
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 gave the signal for the boston tea party?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "Adams" ] }, "style": "rule" }
{ "index": 596, "need_tools_kwargs": true, "question": "who gave the signal for the boston tea party?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "Adams" ] }, "question": "who gave the signal for the boston tea party?" } } } }
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 take my breath away top gun?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "Berlin" ] }, "style": "rule" }
{ "index": 597, "need_tools_kwargs": true, "question": "who sang take my breath away top gun?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "Berlin" ] }, "question": "who sang take my breath away top gun?" } } } }
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 were the 2 presidents to get impeached?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "Bill Clinton", "Andrew Johnson" ] }, "style": "rule" }
{ "index": 598, "need_tools_kwargs": true, "question": "who were the 2 presidents to get impeached?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "Bill Clinton", "Andrew Johnson" ] }, "question": "who were the 2 presidents to get 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: when did bananas become popular in the us?", "role": "user" } ]
fact-reasoning
{ "ground_truth": { "target": [ "1876" ] }, "style": "rule" }
{ "index": 599, "need_tools_kwargs": true, "question": "when did bananas become popular in the us?", "split": "test", "tools_kwargs": { "search": { "create_kwargs": { "data_source": "nq", "ground_truth": { "target": [ "1876" ] }, "question": "when did bananas become popular in the us?" } } } }
null