id
stringlengths 1
12
| question
stringlengths 22
566
| golden_answers
dict | data_source
stringclasses 2
values | prompt
listlengths 1
1
| ability
stringclasses 6
values | reward_model
dict | extra_info
dict | metadata
dict |
|---|---|---|---|---|---|---|---|---|
51464
|
Mitsubishi gave the task of designing the new fighter to Joji Hattori, the result was a single-bay biplane of wooden construction and had a watertight fuselage.
|
{
"evidence": [
"Mitsubishi 1MF9_sentence_4",
"Mitsubishi 1MF9_sentence_5",
"Mitsubishi 1MF9_sentence_6"
],
"label": "SUPPORT"
}
|
feverous
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nMitsubishi gave the task of designing the new fighter to Joji Hattori, the result was a single-bay biplane of wooden construction and had a watertight fuselage.\n",
"role": "user"
}
] |
Search terms not in claim
|
{
"ground_truth": {
"evidence": [
"Mitsubishi 1MF9_sentence_4",
"Mitsubishi 1MF9_sentence_5",
"Mitsubishi 1MF9_sentence_6"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 4384,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.331",
"value": "?search="
},
{
"operation": "back-button-clicked",
"time": "9.233",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "9.423",
"value": "Ndidi Onukwulu"
},
{
"operation": "search",
"time": "78.188",
"value": "Joji Hattori"
},
{
"operation": "Now on",
"time": "79.183",
"value": "Joji Hattori"
},
{
"operation": "search",
"time": "139.973",
"value": "Mitsubishi"
},
{
"operation": "Now on",
"time": "141.122",
"value": "Mitsubishi"
},
{
"operation": "hyperlink",
"time": "223.156",
"value": "List of aircraft by Mitsubishi"
},
{
"operation": "Now on",
"time": "231.166",
"value": "List of aircraft (Mi)#Mitsubishi"
},
{
"operation": "hyperlink",
"time": "248.467",
"value": "1.7 Michigan List of aircraft (Mi)_section_7"
},
{
"operation": "search",
"time": "464.172",
"value": "Single-bay biplane"
},
{
"operation": "Now on",
"time": "466.137",
"value": "undefined"
},
{
"operation": "search",
"time": "558.112",
"value": "Aircraft designer Hattori"
},
{
"operation": "Now on",
"time": "559.072",
"value": "undefined"
},
{
"operation": "hyperlink",
"time": "570.412",
"value": "Mitsubishi 1MF9"
},
{
"operation": "Now on",
"time": "571.216",
"value": "Mitsubishi 1MF9"
},
{
"operation": "Highlighting",
"time": "619.309",
"value": "Mitsubishi 1MF9_sentence_4"
},
{
"operation": "Highlighting",
"time": "630.315",
"value": "Mitsubishi 1MF9_sentence_5"
},
{
"operation": "Highlighting",
"time": "638.354",
"value": "Mitsubishi 1MF9_sentence_6"
},
{
"operation": "finish",
"time": "706.634",
"value": "finish"
}
]
}
|
63933
|
Dau Dayal Joshi is a leader of Bharatiya Janata Part, he is an ayurveda medical practitioner but never became a Member of Parliament.
|
{
"evidence": [
"Dau Dayal Joshi_sentence_1",
"Lok Sabha_sentence_2"
],
"label": "REFUTE"
}
|
feverous
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nDau Dayal Joshi is a leader of Bharatiya Janata Part, he is an ayurveda medical practitioner but never became a Member of Parliament.\n",
"role": "user"
}
] |
Multi-hop Reasoning
|
{
"ground_truth": {
"evidence": [
"Dau Dayal Joshi_sentence_1",
"Lok Sabha_sentence_2"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 4385,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "28.968",
"value": "undefined"
},
{
"operation": "search",
"time": "39.281",
"value": "Dau Dayal Joshi"
},
{
"operation": "Now on",
"time": "41.584",
"value": "Dau Dayal Joshi"
},
{
"operation": "Page search",
"time": "44.588",
"value": "parlia"
},
{
"operation": "Highlighting",
"time": "63.163",
"value": "Dau Dayal Joshi_sentence_2"
},
{
"operation": "Highlighting",
"time": "65.973",
"value": "Dau Dayal Joshi_sentence_3"
},
{
"operation": "hyperlink",
"time": "70.955",
"value": "Lok Sabha"
},
{
"operation": "Now on",
"time": "76.354",
"value": "Dau Dayal Joshi"
},
{
"operation": "Highlighting deleted",
"time": "77.652",
"value": "Dau Dayal Joshi_sentence_3"
},
{
"operation": "Highlighting deleted",
"time": "78.077",
"value": "Dau Dayal Joshi_sentence_2"
},
{
"operation": "Highlighting",
"time": "78.721",
"value": "Dau Dayal Joshi_sentence_1"
},
{
"operation": "hyperlink",
"time": "79.682",
"value": "Lok Sabha"
},
{
"operation": "Now on",
"time": "82.315",
"value": "Parliament of India"
},
{
"operation": "back-button-clicked",
"time": "83.52",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "83.589",
"value": "Lok Sabha"
},
{
"operation": "Highlighting",
"time": "84.799",
"value": "Lok Sabha_sentence_2"
},
{
"operation": "finish",
"time": "90.005",
"value": "finish"
}
]
}
|
exfever_4825
|
Death Note is a series of movie programs that are a staple of television and are broadcast daily and produce a genre of film, television, video game, and theatre in Japan.
|
{
"evidence": [
"Death Note (2015 TV series)",
"Japanese television drama",
"Jidaigeki"
],
"label": "REFUTE"
}
|
exfever
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nDeath Note is a series of movie programs that are a staple of television and are broadcast daily and produce a genre of film, television, video game, and theatre in Japan.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Death Note (2015 TV series)",
"Japanese television drama",
"Jidaigeki"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 4386,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_5142
|
Radiohead is an English rock band formed in a historic market town and civil parish.
|
{
"evidence": [
"Radiohead",
"Abingdon-on-Thames"
],
"label": "SUPPORT"
}
|
exfever
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nRadiohead is an English rock band formed in a historic market town and civil parish.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Radiohead",
"Abingdon-on-Thames"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 4387,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
56529
|
Perth Wildcats won the NBL Championships 10 times in 1990, 1991, 1995, 2000,2010,2014, 2016, 2019, and 2020.
|
{
"evidence": [
"Perth Wildcats_sentence_6"
],
"label": "REFUTE"
}
|
feverous
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nPerth Wildcats won the NBL Championships 10 times in 1990, 1991, 1995, 2000,2010,2014, 2016, 2019, and 2020.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Perth Wildcats_sentence_6"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 4388,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.381",
"value": "?search="
},
{
"operation": "search",
"time": "7.773",
"value": "Perth Wildcats"
},
{
"operation": "Now on",
"time": "10.761",
"value": "Perth Wildcats"
},
{
"operation": "Page search",
"time": "87.338",
"value": "nbl"
},
{
"operation": "Highlighting",
"time": "105.629",
"value": "Perth Wildcats_sentence_6"
},
{
"operation": "finish",
"time": "158.447",
"value": "finish"
}
]
}
|
71074
|
2011–12 FIS Freestyle Skiing World Cup (the thirty third World Cup season) had ten seasons between 2011 and 2012, one of which had the tragic accident of Nick Zoricic.
|
{
"evidence": [
"2011–12 FIS Freestyle Skiing World Cup_sentence_0",
"2011–12 FIS Freestyle Skiing World Cup_sentence_1"
],
"label": "REFUTE"
}
|
feverous
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\n2011–12 FIS Freestyle Skiing World Cup (the thirty third World Cup season) had ten seasons between 2011 and 2012, one of which had the tragic accident of Nick Zoricic.\n",
"role": "user"
}
] |
Entity Disambiguation
|
{
"ground_truth": {
"evidence": [
"2011–12 FIS Freestyle Skiing World Cup_sentence_0",
"2011–12 FIS Freestyle Skiing World Cup_sentence_1"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 4389,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.188",
"value": "?search="
},
{
"operation": "search",
"time": "18.604",
"value": "2011–12 FIS Freestyle Skiing World Cup"
},
{
"operation": "Now on",
"time": "20.339",
"value": "2011–12 FIS Freestyle Skiing World Cup"
},
{
"operation": "Highlighting",
"time": "24.807",
"value": "2011–12 FIS Freestyle Skiing World Cup_sentence_0"
},
{
"operation": "Highlighting",
"time": "51.758",
"value": "2011–12 FIS Freestyle Skiing World Cup_sentence_1"
},
{
"operation": "finish",
"time": "120.528",
"value": "finish"
}
]
}
|
91116
|
The Country of God was published in 1978 by author Witness Lee (Chinese Christian preacher and hymnist belonging to the Christian group) in United States with genre of Christianity.
|
{
"evidence": [
"Witness Lee_sentence_0"
],
"label": "NOT ENOUGH INFO"
}
|
feverous
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nThe Country of God was published in 1978 by author Witness Lee (Chinese Christian preacher and hymnist belonging to the Christian group) in United States with genre of Christianity.\n",
"role": "user"
}
] |
Search terms not in claim
|
{
"ground_truth": {
"evidence": [
"Witness Lee_sentence_0"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 4390,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.398",
"value": "?search="
},
{
"operation": "search",
"time": "11.903",
"value": "Country of God"
},
{
"operation": "Now on",
"time": "15.294",
"value": "undefined"
},
{
"operation": "search",
"time": "26.893",
"value": "the Country of God"
},
{
"operation": "Now on",
"time": "28.578",
"value": "undefined"
},
{
"operation": "search",
"time": "37.204",
"value": "Country of God"
},
{
"operation": "Now on",
"time": "37.797",
"value": "undefined"
},
{
"operation": "search",
"time": "43.459",
"value": "Witness Lee"
},
{
"operation": "Now on",
"time": "44.337",
"value": "Witness Lee"
},
{
"operation": "Page search",
"time": "107.513",
"value": "god"
},
{
"operation": "Page search",
"time": "129.151",
"value": "country of god"
},
{
"operation": "Page search",
"time": "187.189",
"value": "wr"
},
{
"operation": "Page search",
"time": "257.363",
"value": "publish"
},
{
"operation": "Highlighting",
"time": "375.256",
"value": "Witness Lee_sentence_0"
},
{
"operation": "search",
"time": "387.28",
"value": "The Country of God"
},
{
"operation": "Now on",
"time": "399.566",
"value": "Witness Lee"
},
{
"operation": "finish",
"time": "409.103",
"value": "finish"
}
]
}
|
exfever_6605
|
Emmanuel Adebayor is a Togolese professional footballer who previously played for English clubs, including an English basketball club which moved to Etihad Stadium after having played at a football stadium in Moss Side.
|
{
"evidence": [
"Emmanuel Adebayor",
"Manchester City F.C.",
"Maine Road"
],
"label": "REFUTE"
}
|
exfever
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nEmmanuel Adebayor is a Togolese professional footballer who previously played for English clubs, including an English basketball club which moved to Etihad Stadium after having played at a football stadium in Moss Side.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Emmanuel Adebayor",
"Manchester City F.C.",
"Maine Road"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 4391,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
33101
|
Ouibox was a website headquartered in Lexington, Kentucky, the second-largest city in Kentucky, and features a writing tool called OuiWrite.
|
{
"evidence": [
"Ouibox_sentence_0",
"Lexington, Kentucky_sentence_0",
"Ouibox_sentence_1"
],
"label": "SUPPORT"
}
|
feverous
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nOuibox was a website headquartered in Lexington, Kentucky, the second-largest city in Kentucky, and features a writing tool called OuiWrite.\n",
"role": "user"
}
] |
Multi-hop Reasoning
|
{
"ground_truth": {
"evidence": [
"Ouibox_sentence_0",
"Lexington, Kentucky_sentence_0",
"Ouibox_sentence_1"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 4392,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.872",
"value": "?search="
},
{
"operation": "search",
"time": "10.184",
"value": "Ouibox"
},
{
"operation": "Now on",
"time": "10.907",
"value": "Ouibox"
},
{
"operation": "Highlighting",
"time": "59.088",
"value": "Ouibox_sentence_0"
},
{
"operation": "hyperlink",
"time": "62.553",
"value": "Lexington, Kentucky"
},
{
"operation": "Now on",
"time": "63.793",
"value": "Lexington, Kentucky"
},
{
"operation": "Highlighting",
"time": "66.004",
"value": "Lexington, Kentucky_sentence_0"
},
{
"operation": "back-button-clicked",
"time": "75.187",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "75.238",
"value": "Ouibox"
},
{
"operation": "Highlighting",
"time": "78.863",
"value": "Ouibox_sentence_1"
},
{
"operation": "finish",
"time": "103.304",
"value": "finish"
}
]
}
|
90076
|
The XRCC3 protein is one of four paralogs of RAD51, including RAD51B (RAD51L1), RAD51C (RAD51L2), and they each share about 25% amino acid sequence identity with RAD51 and each other.
|
{
"evidence": [
"XRCC3_sentence_4"
],
"label": "REFUTE"
}
|
feverous
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nThe XRCC3 protein is one of four paralogs of RAD51, including RAD51B (RAD51L1), RAD51C (RAD51L2), and they each share about 25% amino acid sequence identity with RAD51 and each other.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"XRCC3_sentence_4"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 4393,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.727",
"value": "?search="
},
{
"operation": "search",
"time": "10.423",
"value": "XRCC3"
},
{
"operation": "Now on",
"time": "11.135",
"value": "XRCC3"
},
{
"operation": "Page search",
"time": "18.925",
"value": "paralog"
},
{
"operation": "Highlighting",
"time": "21.229",
"value": "XRCC3_sentence_4"
},
{
"operation": "finish",
"time": "267.596",
"value": "finish"
}
]
}
|
exfever_7894
|
All My Children was an American long-running television series set in Pine Valley, Pennsylvania, that was characterized by melodrama, ensemble casts, and sentimentality.
|
{
"evidence": [
"All My Children",
"Soap opera"
],
"label": "SUPPORT"
}
|
exfever
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nAll My Children was an American long-running television series set in Pine Valley, Pennsylvania, that was characterized by melodrama, ensemble casts, and sentimentality.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"All My Children",
"Soap opera"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 4394,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_7074
|
The Times was founded by publisher John Walter.
|
{
"evidence": [
"The Times",
"News Corp"
],
"label": "NOT ENOUGH INFO"
}
|
exfever
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nThe Times was founded by publisher John Walter.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"The Times",
"News Corp"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 4395,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
63366
|
Karemlash was the seat of the Nestorian patriarch Denna II and home to Jacobite and Armenian communities in the 11th century.
|
{
"evidence": [
"Karemlash_sentence_5",
"Karemlash_sentence_6"
],
"label": "REFUTE"
}
|
feverous
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nKaremlash was the seat of the Nestorian patriarch Denna II and home to Jacobite and Armenian communities in the 11th century.\n",
"role": "user"
}
] |
Numerical Reasoning
|
{
"ground_truth": {
"evidence": [
"Karemlash_sentence_5",
"Karemlash_sentence_6"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 4396,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.782",
"value": "?search="
},
{
"operation": "search",
"time": "9.391",
"value": "Karemlash"
},
{
"operation": "Now on",
"time": "10.513",
"value": "Karemlash"
},
{
"operation": "Highlighting",
"time": "27.539",
"value": "Karemlash_sentence_4"
},
{
"operation": "Highlighting",
"time": "36.16",
"value": "Karemlash_sentence_6"
},
{
"operation": "Highlighting",
"time": "45.166",
"value": "Karemlash_sentence_5"
},
{
"operation": "Highlighting",
"time": "45.67",
"value": "Karemlash_sentence_6"
},
{
"operation": "finish",
"time": "50.292",
"value": "finish"
}
]
}
|
exfever_5618
|
Sir Richard Starkey is a French musician who occasionally sang lead vocals with the group, usually for one song on each album, including a song by the English rock band the Beatles from their seventh studio album that topped the Record Retailer chart in the UK for seven weeks and the US Billboard Top LPs list for six weeks.
|
{
"evidence": [
"Ringo Starr",
"Yellow Submarine (song)",
"Revolver (Beatles album)"
],
"label": "REFUTE"
}
|
exfever
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nSir Richard Starkey is a French musician who occasionally sang lead vocals with the group, usually for one song on each album, including a song by the English rock band the Beatles from their seventh studio album that topped the Record Retailer chart in the UK for seven weeks and the US Billboard Top LPs list for six weeks.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Ringo Starr",
"Yellow Submarine (song)",
"Revolver (Beatles album)"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 4397,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
37161
|
Arhodomonas aquaeolei causes no known diseases and shares phototropic traits with purple sulfur bacteria.
|
{
"evidence": [
"Arhodomonas aquaeolei_sentence_8"
],
"label": "REFUTE"
}
|
feverous
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nArhodomonas aquaeolei causes no known diseases and shares phototropic traits with purple sulfur bacteria.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Arhodomonas aquaeolei_sentence_8"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 4398,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.994",
"value": "?search="
},
{
"operation": "search",
"time": "8.7",
"value": "Arhodomonas aquaeolei"
},
{
"operation": "Now on",
"time": "9.77",
"value": "Arhodomonas aquaeolei"
},
{
"operation": "Highlighting",
"time": "20.883",
"value": "Arhodomonas aquaeolei_sentence_10"
},
{
"operation": "Highlighting",
"time": "54.084",
"value": "Arhodomonas aquaeolei_sentence_8"
},
{
"operation": "Highlighting deleted",
"time": "55.124",
"value": "Arhodomonas aquaeolei_sentence_10"
},
{
"operation": "finish",
"time": "59.899",
"value": "finish"
}
]
}
|
62512
|
The Iris latistyla has two bract (protective layer) and the flowers are 5 cm in diameter with two pairs of petals.
|
{
"evidence": [
"Iris latistyla_sentence_13",
"Iris latistyla_sentence_14"
],
"label": "NOT ENOUGH INFO"
}
|
feverous
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nThe Iris latistyla has two bract (protective layer) and the flowers are 5 cm in diameter with two pairs of petals.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Iris latistyla_sentence_13",
"Iris latistyla_sentence_14"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 4399,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "2.468",
"value": "?search="
},
{
"operation": "search",
"time": "45.24",
"value": "Iris latistyla"
},
{
"operation": "Now on",
"time": "48.569",
"value": "Iris latistyla"
},
{
"operation": "Highlighting",
"time": "154.028",
"value": "Iris latistyla_sentence_13"
},
{
"operation": "Highlighting",
"time": "154.86",
"value": "Iris latistyla_sentence_14"
},
{
"operation": "finish",
"time": "215.869",
"value": "finish"
}
]
}
|
9347
|
The Palace of Fernán Núñez, being one of the most important buildings of civil character was built in Fernán Núñez, which is in the Córdoba province.
|
{
"evidence": [
"Duke of Fernán Núñez_sentence_3",
"Duke of Fernán Núñez_sentence_4"
],
"label": "SUPPORT"
}
|
feverous
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nThe Palace of Fernán Núñez, being one of the most important buildings of civil character was built in Fernán Núñez, which is in the Córdoba province.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Duke of Fernán Núñez_sentence_3",
"Duke of Fernán Núñez_sentence_4"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 4400,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.933",
"value": "?search="
},
{
"operation": "search",
"time": "31.239",
"value": "Fernán Núñez"
},
{
"operation": "Now on",
"time": "32.173",
"value": "Fernán Núñez"
},
{
"operation": "search",
"time": "42.285",
"value": "Contains...Palace of Fernán Núñez"
},
{
"operation": "Now on",
"time": "43.13",
"value": "undefined"
},
{
"operation": "search",
"time": "62.041",
"value": "Duke of Fernán Núñez"
},
{
"operation": "Now on",
"time": "62.954",
"value": "Duke of Fernán Núñez"
},
{
"operation": "Page search",
"time": "78.964",
"value": "building"
},
{
"operation": "Highlighting",
"time": "81.591",
"value": "Duke of Fernán Núñez_sentence_4"
},
{
"operation": "Highlighting deleted",
"time": "128.004",
"value": "Duke of Fernán Núñez_sentence_4"
},
{
"operation": "Highlighting",
"time": "128.495",
"value": "Duke of Fernán Núñez_sentence_3"
},
{
"operation": "Highlighting",
"time": "128.949",
"value": "Duke of Fernán Núñez_sentence_4"
},
{
"operation": "finish",
"time": "143.358",
"value": "finish"
}
]
}
|
92830
|
Falkland Islands at the 2018 Commonwealth Games in the Gold Coast, Australia and was the Falkland Island's 12th appearance at the Games.
|
{
"evidence": [
"Falkland Islands at the 2018 Commonwealth Games_sentence_1"
],
"label": "REFUTE"
}
|
feverous
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nFalkland Islands at the 2018 Commonwealth Games in the Gold Coast, Australia and was the Falkland Island's 12th appearance at the Games.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Falkland Islands at the 2018 Commonwealth Games_sentence_1"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 4401,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.127",
"value": "?search="
},
{
"operation": "search",
"time": "56.809",
"value": "2018 Commonwealth Games in the Gold Coast"
},
{
"operation": "Now on",
"time": "61.377",
"value": "undefined"
},
{
"operation": "hyperlink",
"time": "92.385",
"value": "Seychelles at the 2018 Commonwealth Games"
},
{
"operation": "Now on",
"time": "93.857",
"value": "Seychelles at the 2018 Commonwealth Games"
},
{
"operation": "Page search",
"time": "128.203",
"value": "Falkland"
},
{
"operation": "back-button-clicked",
"time": "149.164",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "149.969",
"value": "undefined"
},
{
"operation": "search",
"time": "192.49",
"value": "Falkland Islands at the 2018 Commonwealth Games in the Gold Coast"
},
{
"operation": "Now on",
"time": "195.385",
"value": "undefined"
},
{
"operation": "hyperlink",
"time": "199.156",
"value": "Falkland Islands at the 2018 Commonwealth Games"
},
{
"operation": "Now on",
"time": "199.842",
"value": "Falkland Islands at the 2018 Commonwealth Games"
},
{
"operation": "Highlighting",
"time": "224.99",
"value": "Falkland Islands at the 2018 Commonwealth Games_sentence_0"
},
{
"operation": "Highlighting",
"time": "253.091",
"value": "Falkland Islands at the 2018 Commonwealth Games_sentence_1"
},
{
"operation": "Highlighting deleted",
"time": "255.724",
"value": "Falkland Islands at the 2018 Commonwealth Games_sentence_0"
},
{
"operation": "finish",
"time": "273.593",
"value": "finish"
}
]
}
|
14986
|
Folly's first full length album, Insanity Later, includes guest vocalists Joey Southside and Eric Gunderson, and was released on Triple Crown Records, which was founded in 1997 by Fred Feldman.
|
{
"evidence": [
"Insanity Later_sentence_0",
"Insanity Later_sentence_2",
"Triple Crown Records_sentence_0"
],
"label": "SUPPORT"
}
|
feverous
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nFolly's first full length album, Insanity Later, includes guest vocalists Joey Southside and Eric Gunderson, and was released on Triple Crown Records, which was founded in 1997 by Fred Feldman.\n",
"role": "user"
}
] |
Multi-hop Reasoning
|
{
"ground_truth": {
"evidence": [
"Insanity Later_sentence_0",
"Insanity Later_sentence_2",
"Triple Crown Records_sentence_0"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 4402,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.549",
"value": "?search="
},
{
"operation": "search",
"time": "12.323",
"value": "Insanity Later"
},
{
"operation": "Now on",
"time": "13.259",
"value": "Insanity Later"
},
{
"operation": "Highlighting",
"time": "23.072",
"value": "Insanity Later_sentence_0"
},
{
"operation": "Highlighting",
"time": "33.775",
"value": "Insanity Later_sentence_2"
},
{
"operation": "hyperlink",
"time": "39.799",
"value": "Triple Crown Records"
},
{
"operation": "Now on",
"time": "40.345",
"value": "Triple Crown Records"
},
{
"operation": "Highlighting",
"time": "42.52",
"value": "Triple Crown Records_sentence_0"
},
{
"operation": "finish",
"time": "48.079",
"value": "finish"
}
]
}
|
exfever_2228
|
Neymar da Silva Santos Júnior was the wealthiest Brazilian professional footballer.
|
{
"evidence": [
"Neymar",
"Copa do Brasil"
],
"label": "NOT ENOUGH INFO"
}
|
exfever
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nNeymar da Silva Santos Júnior was the wealthiest Brazilian professional footballer.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Neymar",
"Copa do Brasil"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 4403,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
91067
|
Fort Totten station was part of a 36 month trial and was made permanent in 2008.
|
{
"evidence": [
"Fort Totten station_sentence_14"
],
"label": "REFUTE"
}
|
feverous
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nFort Totten station was part of a 36 month trial and was made permanent in 2008.\n",
"role": "user"
}
] |
Numerical Reasoning
|
{
"ground_truth": {
"evidence": [
"Fort Totten station_sentence_14"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 4404,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.059",
"value": "?search="
},
{
"operation": "search",
"time": "9.139",
"value": "Fort Totten"
},
{
"operation": "Now on",
"time": "9.906",
"value": "Fort Totten"
},
{
"operation": "hyperlink",
"time": "17.682",
"value": "Fort Totten (WMATA station)"
},
{
"operation": "Now on",
"time": "18.52",
"value": "Fort Totten station"
},
{
"operation": "Page search",
"time": "23.367",
"value": "trial"
},
{
"operation": "Highlighting",
"time": "26.986",
"value": "Fort Totten station_sentence_14"
},
{
"operation": "Page search",
"time": "42.759",
"value": "permanent"
},
{
"operation": "Page search",
"time": "44.327",
"value": "permanent"
},
{
"operation": "Highlighting deleted",
"time": "56.921",
"value": "Fort Totten station_sentence_14"
},
{
"operation": "Highlighting",
"time": "57.633",
"value": "Fort Totten station_sentence_14"
},
{
"operation": "Page search",
"time": "99.063",
"value": "36"
},
{
"operation": "Page search",
"time": "99.806",
"value": "36"
},
{
"operation": "finish",
"time": "102.057",
"value": "finish"
}
]
}
|
3079
|
Hunter King's longest role in Television Roles is for the TV series The Price is Right while her shortest role in TV series is in The Young and the Restless.
|
{
"evidence": [
"Hunter King_sentence_2",
"Hunter King_sentence_27"
],
"label": "REFUTE"
}
|
feverous
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nHunter King's longest role in Television Roles is for the TV series The Price is Right while her shortest role in TV series is in The Young and the Restless.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Hunter King_sentence_2",
"Hunter King_sentence_27"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 4405,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "search",
"time": "11.263",
"value": "hunter king"
},
{
"operation": "hyperlink",
"time": "29.754",
"value": "2 Career Hunter King_section_1"
},
{
"operation": "Highlighting",
"time": "49.732",
"value": "Hunter King_sentence_27"
},
{
"operation": "Highlighting",
"time": "68.504",
"value": "Hunter King_sentence_2"
},
{
"operation": "Highlighting",
"time": "91.943",
"value": "Hunter King_cell_2_12_0"
},
{
"operation": "Highlighting",
"time": "91.961",
"value": "Hunter King_section_2"
},
{
"operation": "Highlighting",
"time": "91.98",
"value": "Hunter King_header_cell_2_0_0"
},
{
"operation": "Highlighting",
"time": "93.209",
"value": "Hunter King_cell_2_12_1"
},
{
"operation": "Highlighting",
"time": "93.249",
"value": "Hunter King_header_cell_2_0_1"
},
{
"operation": "Highlighting",
"time": "94.498",
"value": "Hunter King_cell_2_12_2"
},
{
"operation": "Highlighting",
"time": "94.545",
"value": "Hunter King_header_cell_2_0_2"
},
{
"operation": "Highlighting",
"time": "95.355",
"value": "Hunter King_cell_2_12_3"
},
{
"operation": "Highlighting",
"time": "95.411",
"value": "Hunter King_header_cell_2_0_3"
},
{
"operation": "Highlighting deleted",
"time": "128.806",
"value": "Hunter King_sentence_27"
},
{
"operation": "Highlighting",
"time": "130.697",
"value": "Hunter King_sentence_27"
},
{
"operation": "Highlighting deleted",
"time": "135.306",
"value": "Hunter King_cell_2_12_1"
},
{
"operation": "Highlighting deleted",
"time": "135.36",
"value": "Hunter King_header_cell_2_0_1"
},
{
"operation": "Highlighting deleted",
"time": "136.27",
"value": "Hunter King_cell_2_12_0"
},
{
"operation": "Highlighting deleted",
"time": "136.319",
"value": "Hunter King_header_cell_2_0_0"
},
{
"operation": "Highlighting deleted",
"time": "137.12",
"value": "Hunter King_cell_2_12_2"
},
{
"operation": "Highlighting deleted",
"time": "137.166",
"value": "Hunter King_header_cell_2_0_2"
},
{
"operation": "Highlighting deleted",
"time": "138.168",
"value": "Hunter King_cell_2_12_3"
},
{
"operation": "Highlighting deleted",
"time": "138.185",
"value": "Hunter King_section_2"
},
{
"operation": "Highlighting deleted",
"time": "138.201",
"value": "Hunter King_header_cell_2_0_3"
},
{
"operation": "finish",
"time": "158.248",
"value": "finish"
}
]
}
|
exfever_7927
|
Marilyn Manson is an American rock band formed by Marilyn Manson and Daisy Berkowitz in the coastal city located in the U.S. state of Florida and is the largest city in Broward County with a population of 232,760.
|
{
"evidence": [
"Marilyn Manson (band)",
"Fort Lauderdale, Florida"
],
"label": "REFUTE"
}
|
exfever
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nMarilyn Manson is an American rock band formed by Marilyn Manson and Daisy Berkowitz in the coastal city located in the U.S. state of Florida and is the largest city in Broward County with a population of 232,760.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Marilyn Manson (band)",
"Fort Lauderdale, Florida"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 4406,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
58009
|
Elliot Kebbie plays as a defender or midfielder and is a former England U-16's international, a team that represents England in association football at an under-16 age level.
|
{
"evidence": [
"Elliot Kebbie_sentence_0",
"Elliot Kebbie_sentence_2",
"England national under-16 football team_sentence_0"
],
"label": "SUPPORT"
}
|
feverous
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nElliot Kebbie plays as a defender or midfielder and is a former England U-16's international, a team that represents England in association football at an under-16 age level.\n",
"role": "user"
}
] |
Multi-hop Reasoning
|
{
"ground_truth": {
"evidence": [
"Elliot Kebbie_sentence_0",
"Elliot Kebbie_sentence_2",
"England national under-16 football team_sentence_0"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 4407,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.087",
"value": "?search="
},
{
"operation": "search",
"time": "7.281",
"value": "Elliot Kebbie"
},
{
"operation": "Now on",
"time": "8.115",
"value": "Elliot Kebbie"
},
{
"operation": "Highlighting",
"time": "16.116",
"value": "Elliot Kebbie_sentence_0"
},
{
"operation": "Highlighting",
"time": "46.716",
"value": "Elliot Kebbie_sentence_2"
},
{
"operation": "hyperlink",
"time": "53.258",
"value": "England U-16"
},
{
"operation": "Now on",
"time": "54.073",
"value": "England national under-16 football team"
},
{
"operation": "Highlighting",
"time": "60.217",
"value": "England national under-16 football team_sentence_0"
},
{
"operation": "finish",
"time": "79.599",
"value": "finish"
}
]
}
|
16390
|
OT refers to overtime/shootout losses, while PTS refers to post-traumatic stress.
|
{
"evidence": [
"Post-traumatic stress disorder_sentence_4",
"2004–05 AHL season_sentence_12"
],
"label": "REFUTE"
}
|
feverous
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nOT refers to overtime/shootout losses, while PTS refers to post-traumatic stress.\n",
"role": "user"
}
] |
Multi-hop Reasoning
|
{
"ground_truth": {
"evidence": [
"Post-traumatic stress disorder_sentence_4",
"2004–05 AHL season_sentence_12"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 4408,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.632",
"value": "?search="
},
{
"operation": "search",
"time": "9.263",
"value": "OT refers to overtime/shootout losses,"
},
{
"operation": "Now on",
"time": "11.29",
"value": "undefined"
},
{
"operation": "search",
"time": "61.611",
"value": "OT"
},
{
"operation": "Now on",
"time": "62.626",
"value": "OT"
},
{
"operation": "search",
"time": "145.239",
"value": "overtime/shootout losses"
},
{
"operation": "Now on",
"time": "146.35",
"value": "undefined"
},
{
"operation": "search",
"time": "185.062",
"value": "PTS"
},
{
"operation": "Now on",
"time": "186.448",
"value": "PTS"
},
{
"operation": "Highlighting",
"time": "221.316",
"value": "PTS_item_4_12"
},
{
"operation": "Highlighting",
"time": "221.318",
"value": "PTS_title"
},
{
"operation": "Highlighting",
"time": "221.318",
"value": "PTS_section_5"
},
{
"operation": "Highlighting",
"time": "221.319",
"value": "PTS_section_6"
},
{
"operation": "Highlighting",
"time": "221.825",
"value": "PTS_item_4_11"
},
{
"operation": "Highlighting",
"time": "222.273",
"value": "PTS_item_4_10"
},
{
"operation": "Highlighting",
"time": "222.737",
"value": "PTS_item_4_9"
},
{
"operation": "search",
"time": "236.602",
"value": "post-traumatic stress"
},
{
"operation": "Now on",
"time": "239.467",
"value": "Post-traumatic stress disorder"
},
{
"operation": "Highlighting",
"time": "250.354",
"value": "Post-traumatic stress disorder_sentence_4"
},
{
"operation": "hyperlink",
"time": "257.633",
"value": "PTSD (disambiguation)"
},
{
"operation": "Now on",
"time": "258.632",
"value": "PTSD (disambiguation)"
},
{
"operation": "back-button-clicked",
"time": "271.705",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "274.251",
"value": "Post-traumatic stress disorder"
},
{
"operation": "search",
"time": "292.553",
"value": "overtime losses"
},
{
"operation": "Now on",
"time": "293.512",
"value": "undefined"
},
{
"operation": "hyperlink",
"time": "299.895",
"value": "HYS The Hague"
},
{
"operation": "Now on",
"time": "300.749",
"value": "HYS The Hague"
},
{
"operation": "back-button-clicked",
"time": "303.84",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "304.716",
"value": "undefined"
},
{
"operation": "hyperlink",
"time": "311.391",
"value": "2004–05 AHL season"
},
{
"operation": "Now on",
"time": "313.367",
"value": "2004–05 AHL season"
},
{
"operation": "Page search",
"time": "319.918",
"value": "OT"
},
{
"operation": "Highlighting",
"time": "327.953",
"value": "2004–05 AHL season_sentence_12"
},
{
"operation": "finish",
"time": "346.474",
"value": "finish"
}
]
}
|
exfever_1509
|
My Own Private Idaho is a 1991 American feature film that was produced outside of the major film studio system.
|
{
"evidence": [
"My Own Private Idaho",
"Independent film"
],
"label": "SUPPORT"
}
|
exfever
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nMy Own Private Idaho is a 1991 American feature film that was produced outside of the major film studio system.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"My Own Private Idaho",
"Independent film"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 4409,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
41796
|
Lizanne Bussières represent Malaysia in different competitions like Olympic games and SEA Games from year 1986 to 2204 and got the 1st place in Pittsburgh Marathon.
|
{
"evidence": [
"Lizanne Bussières_sentence_0"
],
"label": "REFUTE"
}
|
feverous
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nLizanne Bussières represent Malaysia in different competitions like Olympic games and SEA Games from year 1986 to 2204 and got the 1st place in Pittsburgh Marathon.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Lizanne Bussières_sentence_0"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 4410,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.071",
"value": "?search="
},
{
"operation": "search",
"time": "12.863",
"value": "Lizanne Bussières"
},
{
"operation": "Now on",
"time": "13.748",
"value": "Lizanne Bussières"
},
{
"operation": "Highlighting",
"time": "24.321",
"value": "Lizanne Bussières_sentence_0"
},
{
"operation": "hyperlink",
"time": "64.275",
"value": "Commonwealth Games"
},
{
"operation": "Now on",
"time": "65.098",
"value": "Athletics at the 1986 Commonwealth Games"
},
{
"operation": "back-button-clicked",
"time": "79.036",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "79.087",
"value": "Lizanne Bussières"
},
{
"operation": "hyperlink",
"time": "86.531",
"value": "2"
},
{
"operation": "Now on",
"time": "86.965",
"value": "Athletics at the 1986 Commonwealth Games – Women's marathon"
},
{
"operation": "back-button-clicked",
"time": "94.188",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "94.231",
"value": "Lizanne Bussières"
},
{
"operation": "finish",
"time": "105.11",
"value": "finish"
}
]
}
|
37019
|
Hyde Park Academy High School's demographic started to change during the mid-1940s and in 1967, the school's students were 97% African-American.
|
{
"evidence": [
"Hyde Park Academy High School_sentence_11",
"Hyde Park Academy High School_sentence_15"
],
"label": "SUPPORT"
}
|
feverous
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nHyde Park Academy High School's demographic started to change during the mid-1940s and in 1967, the school's students were 97% African-American.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Hyde Park Academy High School_sentence_11",
"Hyde Park Academy High School_sentence_15"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 4411,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.784",
"value": "?search="
},
{
"operation": "search",
"time": "12.04",
"value": "Hyde Park Academy High School"
},
{
"operation": "Now on",
"time": "12.878",
"value": "Hyde Park Academy High School"
},
{
"operation": "Page search",
"time": "26.031",
"value": "1967"
},
{
"operation": "Page search",
"time": "48.369",
"value": "demographic"
},
{
"operation": "Highlighting",
"time": "54.432",
"value": "Hyde Park Academy High School_sentence_11"
},
{
"operation": "Page search",
"time": "61.19",
"value": "1967"
},
{
"operation": "Highlighting",
"time": "66.384",
"value": "Hyde Park Academy High School_sentence_15"
},
{
"operation": "finish",
"time": "215.165",
"value": "finish"
}
]
}
|
exfever_4882
|
American Gods is a fantasy novel by an Italian author of comic books who has won a literary award given by the Association for Library Service to Children to the author of "the most distinguished contribution to American literature for children".
|
{
"evidence": [
"American Gods",
"Neil Gaiman",
"Newbery Medal"
],
"label": "REFUTE"
}
|
exfever
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nAmerican Gods is a fantasy novel by an Italian author of comic books who has won a literary award given by the Association for Library Service to Children to the author of \"the most distinguished contribution to American literature for children\".\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"American Gods",
"Neil Gaiman",
"Newbery Medal"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 4412,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_3703
|
Champion is a 1949 American film noir sports film directed by Mark Robson whose genre is in a category of narrative fiction intended to be more serious than humorous in tone.
|
{
"evidence": [
"Champion (1949 film)",
"Drama (film and television)"
],
"label": "SUPPORT"
}
|
exfever
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nChampion is a 1949 American film noir sports film directed by Mark Robson whose genre is in a category of narrative fiction intended to be more serious than humorous in tone.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Champion (1949 film)",
"Drama (film and television)"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 4413,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
84430
|
The Department of the Air Force is headed by the Secretary of the Air Force, a civilian appointed by the president, by and with the advice and consent of the Senate.
|
{
"evidence": [
"United States Department of the Air Force_sentence_3",
"United States Secretary of the Air Force_sentence_1"
],
"label": "SUPPORT"
}
|
feverous
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nThe Department of the Air Force is headed by the Secretary of the Air Force, a civilian appointed by the president, by and with the advice and consent of the Senate.\n",
"role": "user"
}
] |
Multi-hop Reasoning
|
{
"ground_truth": {
"evidence": [
"United States Department of the Air Force_sentence_3",
"United States Secretary of the Air Force_sentence_1"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 4414,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.989",
"value": "?search="
},
{
"operation": "search",
"time": "13.665",
"value": "Contains...The Department of the Air Force"
},
{
"operation": "Now on",
"time": "15.837",
"value": "undefined"
},
{
"operation": "search",
"time": "27.968",
"value": "Department of the Air Force"
},
{
"operation": "Now on",
"time": "28.999",
"value": "United States Department of the Air Force"
},
{
"operation": "Page search",
"time": "36.753",
"value": "secretary"
},
{
"operation": "Highlighting",
"time": "42.672",
"value": "United States Department of the Air Force_sentence_3"
},
{
"operation": "Page search",
"time": "63.921",
"value": "senate"
},
{
"operation": "hyperlink",
"time": "80.304",
"value": "Secretary of the Air Force"
},
{
"operation": "Now on",
"time": "81.375",
"value": "United States Secretary of the Air Force"
},
{
"operation": "Page search",
"time": "86.166",
"value": "senate"
},
{
"operation": "Highlighting",
"time": "90.624",
"value": "United States Secretary of the Air Force_sentence_1"
},
{
"operation": "finish",
"time": "328.511",
"value": "finish"
}
]
}
|
exfever_4508
|
Jim Parsons is an actor who played a fictional character in the spinoff series of The Big Bang Theory.
|
{
"evidence": [
"Jim Parsons",
"Sheldon Cooper"
],
"label": "SUPPORT"
}
|
exfever
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nJim Parsons is an actor who played a fictional character in the spinoff series of The Big Bang Theory.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Jim Parsons",
"Sheldon Cooper"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 4415,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_6775
|
"Wild Ones" is a song by Canadian rapper Flo Rida from his fourth studio album, Wild Ones, released on July 3, 2012.
|
{
"evidence": [
"Wild Ones (song)",
"Wild Ones (Flo Rida album)"
],
"label": "REFUTE"
}
|
exfever
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\n\"Wild Ones\" is a song by Canadian rapper Flo Rida from his fourth studio album, Wild Ones, released on July 3, 2012.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Wild Ones (song)",
"Wild Ones (Flo Rida album)"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 4416,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
7974
|
Ulane Bonnel (born on September 22, 1918) served in the congressional liaison unit before leaving in 1950 after she met Paul-Henri Bonnei who she later married.
|
{
"evidence": [
"Ulane Bonnel_sentence_0",
"Ulane Bonnel_sentence_6",
"Ulane Bonnel_sentence_5"
],
"label": "REFUTE"
}
|
feverous
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nUlane Bonnel (born on September 22, 1918) served in the congressional liaison unit before leaving in 1950 after she met Paul-Henri Bonnei who she later married.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Ulane Bonnel_sentence_0",
"Ulane Bonnel_sentence_6",
"Ulane Bonnel_sentence_5"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 4417,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.099",
"value": "?search="
},
{
"operation": "search",
"time": "121.641",
"value": "Ulane Bonnel"
},
{
"operation": "Now on",
"time": "122.553",
"value": "Ulane Bonnel"
},
{
"operation": "Highlighting",
"time": "127.544",
"value": "Ulane Bonnel_sentence_0"
},
{
"operation": "Highlighting",
"time": "154.664",
"value": "Ulane Bonnel_sentence_4"
},
{
"operation": "Highlighting",
"time": "190.963",
"value": "Ulane Bonnel_sentence_6"
},
{
"operation": "Highlighting",
"time": "220.157",
"value": "Ulane Bonnel_sentence_5"
},
{
"operation": "Highlighting deleted",
"time": "224.661",
"value": "Ulane Bonnel_sentence_4"
},
{
"operation": "finish",
"time": "232.369",
"value": "finish"
}
]
}
|
exfever_6026
|
Slash is best known as the lead guitarist of an American hard rock band that was disbanded in 1985.
|
{
"evidence": [
"Slash (musician)",
"Guns N' Roses"
],
"label": "REFUTE"
}
|
exfever
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nSlash is best known as the lead guitarist of an American hard rock band that was disbanded in 1985.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Slash (musician)",
"Guns N' Roses"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 4418,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_4866
|
Abel Makkonen Tesfaye is a Canadian singer-songwriter who received 15 awards that are often referred to as the Canadian equivalent of the Brit Awards in Great Britain or the Grammy Awards given in a country primarily located in North America that consists of 50 states.
|
{
"evidence": [
"The Weeknd",
"Juno Award",
"United States"
],
"label": "NOT ENOUGH INFO"
}
|
exfever
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nAbel Makkonen Tesfaye is a Canadian singer-songwriter who received 15 awards that are often referred to as the Canadian equivalent of the Brit Awards in Great Britain or the Grammy Awards given in a country primarily located in North America that consists of 50 states.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"The Weeknd",
"Juno Award",
"United States"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 4419,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
82455
|
On June 23, 2010, was the first season premiere of the twelve weekly episodes of Black Panther in iTunes .
|
{
"evidence": [
"Black panther (disambiguation)_sentence_1"
],
"label": "NOT ENOUGH INFO"
}
|
feverous
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nOn June 23, 2010, was the first season premiere of the twelve weekly episodes of Black Panther in iTunes .\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Black panther (disambiguation)_sentence_1"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 4420,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.072",
"value": "?search="
},
{
"operation": "search",
"time": "14.593",
"value": "Contains...weekly episodes of Black Panther in iTunes"
},
{
"operation": "Now on",
"time": "16.687",
"value": "undefined"
},
{
"operation": "search",
"time": "33.641",
"value": "Contains...Black Panther in iTunes"
},
{
"operation": "Now on",
"time": "34.919",
"value": "undefined"
},
{
"operation": "search",
"time": "65.101",
"value": "Black Panther (soundtrack)"
},
{
"operation": "Now on",
"time": "67.056",
"value": "Black Panther (soundtrack)"
},
{
"operation": "Page search",
"time": "127.709",
"value": "episodes"
},
{
"operation": "search",
"time": "148.101",
"value": "Contains...Panther in iTunes"
},
{
"operation": "Now on",
"time": "149.034",
"value": "undefined"
},
{
"operation": "hyperlink",
"time": "167.005",
"value": "Steel Panther"
},
{
"operation": "Now on",
"time": "167.864",
"value": "Steel Panther"
},
{
"operation": "Page search",
"time": "183.432",
"value": "episodes"
},
{
"operation": "search",
"time": "228.857",
"value": "Contains...weekly episodes in iTunes"
},
{
"operation": "Now on",
"time": "230.471",
"value": "undefined"
},
{
"operation": "search",
"time": "263.663",
"value": "Contains...season premiere of the twelve weekly episodes of"
},
{
"operation": "Now on",
"time": "265.389",
"value": "undefined"
},
{
"operation": "hyperlink",
"time": "284.328",
"value": "Lost (TV series)"
},
{
"operation": "Now on",
"time": "286.047",
"value": "Lost (TV series)"
},
{
"operation": "search",
"time": "333.494",
"value": "Contains...weekly episodes of Black Panther"
},
{
"operation": "Now on",
"time": "334.762",
"value": "undefined"
},
{
"operation": "search",
"time": "362.775",
"value": "Black Panther (film)"
},
{
"operation": "Now on",
"time": "364.998",
"value": "Black Panther (film)"
},
{
"operation": "search",
"time": "402.303",
"value": "Contains... weekly episodes of Black Panther in iTunes"
},
{
"operation": "Now on",
"time": "403.164",
"value": "undefined"
},
{
"operation": "search",
"time": "453.513",
"value": "Contains...June 23, 2010 Panther"
},
{
"operation": "Now on",
"time": "455.66",
"value": "undefined"
},
{
"operation": "hyperlink",
"time": "476.11",
"value": "Marvel Comics"
},
{
"operation": "Now on",
"time": "477.527",
"value": "Marvel Comics"
},
{
"operation": "search",
"time": "496.914",
"value": "Black panther (disambiguation)"
},
{
"operation": "Now on",
"time": "497.837",
"value": "Black panther (disambiguation)"
},
{
"operation": "hyperlink",
"time": "507.571",
"value": "2.3 Other uses in arts and entertainment Black panther (disambiguation)_section_4"
},
{
"operation": "hyperlink",
"time": "545.992",
"value": "Black Panther (band)"
},
{
"operation": "Now on",
"time": "547.225",
"value": "Black Panther (band)"
},
{
"operation": "Page search",
"time": "555.142",
"value": "iTunes"
},
{
"operation": "Now on",
"time": "592.18",
"value": "Black Panther (band)"
},
{
"operation": "search",
"time": "608.944",
"value": "Black Panther (TV series)"
},
{
"operation": "Now on",
"time": "610.256",
"value": "Black Panther (TV series)"
},
{
"operation": "back-button-clicked",
"time": "636.101",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "636.426",
"value": "Black Panther (band)"
},
{
"operation": "search",
"time": "643.786",
"value": "Black panther (disambiguation)"
},
{
"operation": "Now on",
"time": "644.768",
"value": "Black panther (disambiguation)"
},
{
"operation": "Highlighting",
"time": "650.391",
"value": "Black panther (disambiguation)_sentence_1"
},
{
"operation": "finish",
"time": "656.627",
"value": "finish"
}
]
}
|
56667
|
Diodia virginiana is a plant species in the Rubiaceae which is native to Cuba, Nicaragua, Mexico, Connecticut and the south-central and southeastern United States.
|
{
"evidence": [
"Diodia virginiana_sentence_0",
"Diodia virginiana_sentence_5"
],
"label": "SUPPORT"
}
|
feverous
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nDiodia virginiana is a plant species in the Rubiaceae which is native to Cuba, Nicaragua, Mexico, Connecticut and the south-central and southeastern United States.\n",
"role": "user"
}
] |
Entity Disambiguation
|
{
"ground_truth": {
"evidence": [
"Diodia virginiana_sentence_0",
"Diodia virginiana_sentence_5"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 4421,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.223",
"value": "?search="
},
{
"operation": "search",
"time": "12.533",
"value": "Diodia virginiana"
},
{
"operation": "Now on",
"time": "13.933",
"value": "Diodia virginiana"
},
{
"operation": "Highlighting",
"time": "25.079",
"value": "Diodia virginiana_sentence_0"
},
{
"operation": "Highlighting",
"time": "33.798",
"value": "Diodia virginiana_sentence_5"
},
{
"operation": "finish",
"time": "98.677",
"value": "finish"
}
]
}
|
exfever_6880
|
Wyatt Earp was an American lawman who took part in the famous thirty-second shootout between lawmen led by Virgil Earp and members of the Cowboys.
|
{
"evidence": [
"Wyatt Earp",
"Gunfight at the O.K. Corral"
],
"label": "SUPPORT"
}
|
exfever
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nWyatt Earp was an American lawman who took part in the famous thirty-second shootout between lawmen led by Virgil Earp and members of the Cowboys.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Wyatt Earp",
"Gunfight at the O.K. Corral"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 4422,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_6657
|
The Nile is a major north-flowing river whose drainage basin covers eleven countries, including a landlocked country in East Africa.
|
{
"evidence": [
"Nile",
"Uganda"
],
"label": "SUPPORT"
}
|
exfever
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nThe Nile is a major north-flowing river whose drainage basin covers eleven countries, including a landlocked country in East Africa.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Nile",
"Uganda"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 4423,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_702
|
Jenna Dewan was the youngest backup dancer for Janet Jackson.
|
{
"evidence": [
"Jenna Dewan",
"NBC"
],
"label": "NOT ENOUGH INFO"
}
|
exfever
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nJenna Dewan was the youngest backup dancer for Janet Jackson.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Jenna Dewan",
"NBC"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 4424,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_8518
|
The Lamniformes include more unusual representatives, such as the living fossil.
|
{
"evidence": [
"Lamniformes",
"Goblin shark"
],
"label": "SUPPORT"
}
|
exfever
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nThe Lamniformes include more unusual representatives, such as the living fossil.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Lamniformes",
"Goblin shark"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 4425,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_4275
|
Fergie Duhamel is an American singer who vocally incorporates a musical form of vocal delivery into some of her performances with precursors to modern rap including a music genre that originated in the African-American communities of New Orleans, Louisiana.
|
{
"evidence": [
"Fergie (singer)",
"Rapping",
"Jazz"
],
"label": "SUPPORT"
}
|
exfever
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nFergie Duhamel is an American singer who vocally incorporates a musical form of vocal delivery into some of her performances with precursors to modern rap including a music genre that originated in the African-American communities of New Orleans, Louisiana.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Fergie (singer)",
"Rapping",
"Jazz"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 4426,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
56945
|
Late Night Berlin, premiered on March 12, 2018 and hosted by Klaas Heufer-Umlauf, is the only late night comedy show that is really recorded in the said location.
|
{
"evidence": [
"Late Night Berlin_sentence_10",
"Late Night Berlin_sentence_11",
"Late Night Berlin_sentence_0",
"Late Night Berlin_sentence_1"
],
"label": "NOT ENOUGH INFO"
}
|
feverous
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nLate Night Berlin, premiered on March 12, 2018 and hosted by Klaas Heufer-Umlauf, is the only late night comedy show that is really recorded in the said location.\n",
"role": "user"
}
] |
Multi-hop Reasoning
|
{
"ground_truth": {
"evidence": [
"Late Night Berlin_sentence_10",
"Late Night Berlin_sentence_11",
"Late Night Berlin_sentence_0",
"Late Night Berlin_sentence_1"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 4427,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.608",
"value": "?search="
},
{
"operation": "search",
"time": "39.357",
"value": "Late Night Berlin"
},
{
"operation": "Now on",
"time": "40.516",
"value": "Late Night Berlin"
},
{
"operation": "Page search",
"time": "62.148",
"value": "only"
},
{
"operation": "Highlighting",
"time": "109.21",
"value": "Late Night Berlin_sentence_10"
},
{
"operation": "Highlighting",
"time": "110.122",
"value": "Late Night Berlin_sentence_11"
},
{
"operation": "Highlighting",
"time": "150.473",
"value": "Late Night Berlin_sentence_0"
},
{
"operation": "Highlighting",
"time": "154.544",
"value": "Late Night Berlin_sentence_1"
},
{
"operation": "finish",
"time": "167.6",
"value": "finish"
}
]
}
|
exfever_6489
|
Los Angeles is the most populous city in the United States.
|
{
"evidence": [
"Chicago",
"U.S. state"
],
"label": "NOT ENOUGH INFO"
}
|
exfever
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nLos Angeles is the most populous city in the United States.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Chicago",
"U.S. state"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 4428,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_3003
|
Jeb Bush is an American politician and businessman which is the second son of the American politician, diplomat, and businessman who served as the 42nd president of the United States from 1989 to 1993.
|
{
"evidence": [
"Jeb Bush",
"George H. W. Bush"
],
"label": "REFUTE"
}
|
exfever
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nJeb Bush is an American politician and businessman which is the second son of the American politician, diplomat, and businessman who served as the 42nd president of the United States from 1989 to 1993.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Jeb Bush",
"George H. W. Bush"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 4429,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
57222
|
Gary Hunt of Great Britain won the high diving event at the 2015 World Aquatics Championships – Men held on August 3 and 5.
|
{
"evidence": [
"Gary Hunt_sentence_0"
],
"label": "NOT ENOUGH INFO"
}
|
feverous
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nGary Hunt of Great Britain won the high diving event at the 2015 World Aquatics Championships – Men held on August 3 and 5.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Gary Hunt_sentence_0"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 4430,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.108",
"value": "?search="
},
{
"operation": "search",
"time": "15.63",
"value": "Gary Hunt"
},
{
"operation": "Now on",
"time": "16.47",
"value": "Gary Hunt"
},
{
"operation": "Page search",
"time": "34.871",
"value": "2015"
},
{
"operation": "Highlighting",
"time": "57.103",
"value": "Gary Hunt_sentence_0"
},
{
"operation": "Page search",
"time": "94.186",
"value": "great britain"
},
{
"operation": "Page search",
"time": "114.491",
"value": "born"
},
{
"operation": "finish",
"time": "129.368",
"value": "finish"
}
]
}
|
exfever_1484
|
Pulp Fiction is a 1994 American black comedy film whose genre is inspired by and analogous to the crime fiction literary genre that may overlap and combine with many other genres such as a genre of film that revolves around the solution of a problem or a crime.
|
{
"evidence": [
"Pulp Fiction",
"Crime film",
"Mystery film"
],
"label": "SUPPORT"
}
|
exfever
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nPulp Fiction is a 1994 American black comedy film whose genre is inspired by and analogous to the crime fiction literary genre that may overlap and combine with many other genres such as a genre of film that revolves around the solution of a problem or a crime.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Pulp Fiction",
"Crime film",
"Mystery film"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 4431,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_5714
|
Buzz Aldrin is a Canadian former astronaut born in a borough in Essex County, New Jersey, United States.
|
{
"evidence": [
"Buzz Aldrin",
"Glen Ridge, New Jersey"
],
"label": "REFUTE"
}
|
exfever
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nBuzz Aldrin is a Canadian former astronaut born in a borough in Essex County, New Jersey, United States.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Buzz Aldrin",
"Glen Ridge, New Jersey"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 4432,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_7721
|
Karl Anthony Malone spent his first 20 seasons in the NBA?with the Utah Jazz and formed a formidable duo with his teammate,?regarded as one of the greatest point guards and passers of all time.
|
{
"evidence": [
"Karl Malone",
"John Stockton"
],
"label": "REFUTE"
}
|
exfever
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nKarl Anthony Malone spent his first 20 seasons in the NBA?with the Utah Jazz and formed a formidable duo with his teammate,?regarded as one of the greatest point guards and passers of all time.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Karl Malone",
"John Stockton"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 4433,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
93560
|
Ema Twumasi Twumasi was selected 11th overall in the 2018 MLS SuperDraft by FC Dallas, he was loaned to United Soccer League but never played a match for them.
|
{
"evidence": [
"Ema Twumasi_sentence_5",
"Ema Twumasi_sentence_6"
],
"label": "REFUTE"
}
|
feverous
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nEma Twumasi Twumasi was selected 11th overall in the 2018 MLS SuperDraft by FC Dallas, he was loaned to United Soccer League but never played a match for them.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Ema Twumasi_sentence_5",
"Ema Twumasi_sentence_6"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 4434,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "4.291",
"value": "undefined"
},
{
"operation": "search",
"time": "378.305",
"value": "Ema Twumasi Twumasi"
},
{
"operation": "Now on",
"time": "381.356",
"value": "undefined"
},
{
"operation": "hyperlink",
"time": "386.641",
"value": "Ema Twumasi"
},
{
"operation": "Now on",
"time": "387.788",
"value": "Ema Twumasi"
},
{
"operation": "Page search",
"time": "434.679",
"value": "SuperDraft"
},
{
"operation": "Highlighting",
"time": "445.25",
"value": "Ema Twumasi_sentence_4"
},
{
"operation": "Highlighting",
"time": "463.666",
"value": "Ema Twumasi_sentence_5"
},
{
"operation": "Highlighting",
"time": "468.685",
"value": "Ema Twumasi_sentence_6"
},
{
"operation": "Highlighting deleted",
"time": "472.955",
"value": "Ema Twumasi_sentence_4"
},
{
"operation": "page-search-reset",
"time": "476.73",
"value": "page-search-reset"
},
{
"operation": "finish",
"time": "496.665",
"value": "finish"
}
]
}
|
exfever_4817
|
Muhammad was an Arab religious and political leader and the founder of the second-largest religion, whose central text, the Quran, is considered by Muslims to be the word of God.
|
{
"evidence": [
"Muhammad",
"Islam"
],
"label": "SUPPORT"
}
|
exfever
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nMuhammad was an Arab religious and political leader and the founder of the second-largest religion, whose central text, the Quran, is considered by Muslims to be the word of God.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Muhammad",
"Islam"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 4435,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
23349
|
In the Ramsar Convention, an international treaty or formal and binding written agreement, the members of the Standing Committee, an intersessional executive body, are elected in the meetings of the Conference of the Contracting Parties to serve for three years.
|
{
"evidence": [
"Ramsar Convention_sentence_0",
"Treaty_sentence_3",
"Ramsar Convention_sentence_3"
],
"label": "NOT ENOUGH INFO"
}
|
feverous
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nIn the Ramsar Convention, an international treaty or formal and binding written agreement, the members of the Standing Committee, an intersessional executive body, are elected in the meetings of the Conference of the Contracting Parties to serve for three years.\n",
"role": "user"
}
] |
Multi-hop Reasoning
|
{
"ground_truth": {
"evidence": [
"Ramsar Convention_sentence_0",
"Treaty_sentence_3",
"Ramsar Convention_sentence_3"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 4436,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.37",
"value": "?search="
},
{
"operation": "search",
"time": "15.53",
"value": "Ramsar Convention"
},
{
"operation": "Now on",
"time": "16.65",
"value": "Ramsar Convention"
},
{
"operation": "Page search",
"time": "96.31",
"value": "binding"
},
{
"operation": "hyperlink",
"time": "112.654",
"value": "treaty"
},
{
"operation": "Now on",
"time": "113.595",
"value": "Treaty"
},
{
"operation": "back-button-clicked",
"time": "138.09",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "138.664",
"value": "Ramsar Convention"
},
{
"operation": "Highlighting",
"time": "184.542",
"value": "Ramsar Convention_sentence_0"
},
{
"operation": "hyperlink",
"time": "196.657",
"value": "treaty"
},
{
"operation": "Now on",
"time": "197.712",
"value": "Treaty"
},
{
"operation": "Highlighting",
"time": "202.728",
"value": "Treaty_sentence_3"
},
{
"operation": "Page search",
"time": "236.015",
"value": "standing"
},
{
"operation": "Page search",
"time": "279.356",
"value": "standing"
},
{
"operation": "Page search",
"time": "281.235",
"value": "standing"
},
{
"operation": "Page search",
"time": "281.522",
"value": "standing"
},
{
"operation": "Page search",
"time": "321.017",
"value": "committee"
},
{
"operation": "Highlighting",
"time": "336.395",
"value": "Treaty_item_1_12"
},
{
"operation": "Highlighting",
"time": "336.405",
"value": "Treaty_section_26"
},
{
"operation": "Highlighting deleted",
"time": "339.1",
"value": "Treaty_item_1_12"
},
{
"operation": "Highlighting deleted",
"time": "339.109",
"value": "Treaty_section_26"
},
{
"operation": "Page search",
"time": "361.793",
"value": "serve"
},
{
"operation": "Page search",
"time": "412.565",
"value": "Conference of the Contracting Parties"
},
{
"operation": "search",
"time": "436.56",
"value": "Conference of the Contracting Parties"
},
{
"operation": "Now on",
"time": "439.4",
"value": "undefined"
},
{
"operation": "hyperlink",
"time": "450.43",
"value": "Ramsar Convention"
},
{
"operation": "Now on",
"time": "451.483",
"value": "Ramsar Convention"
},
{
"operation": "Page search",
"time": "457.331",
"value": "Conference of the Contracting Parties"
},
{
"operation": "Highlighting",
"time": "472.826",
"value": "Ramsar Convention_sentence_3"
},
{
"operation": "Highlighting deleted",
"time": "557.577",
"value": "Ramsar Convention_sentence_3"
},
{
"operation": "Highlighting",
"time": "558.798",
"value": "Ramsar Convention_sentence_0"
},
{
"operation": "hyperlink",
"time": "559.92",
"value": "treaty"
},
{
"operation": "Now on",
"time": "561.025",
"value": "Treaty"
},
{
"operation": "Highlighting deleted",
"time": "562.924",
"value": "Treaty_sentence_3"
},
{
"operation": "Highlighting",
"time": "564.064",
"value": "Treaty_sentence_3"
},
{
"operation": "back-button-clicked",
"time": "566.318",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "566.86",
"value": "Ramsar Convention"
},
{
"operation": "Highlighting",
"time": "570.632",
"value": "Ramsar Convention_sentence_3"
},
{
"operation": "finish",
"time": "580.48",
"value": "finish"
}
]
}
|
49441
|
Virginia Madsen is an American actress and film producer who made her film debut in Class, which is a comedy-drama film.
|
{
"evidence": [
"Virginia Madsen_sentence_0",
"Virginia Madsen_sentence_1",
"Class (film)_sentence_0"
],
"label": "SUPPORT"
}
|
feverous
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nVirginia Madsen is an American actress and film producer who made her film debut in Class, which is a comedy-drama film.\n",
"role": "user"
}
] |
Multi-hop Reasoning
|
{
"ground_truth": {
"evidence": [
"Virginia Madsen_sentence_0",
"Virginia Madsen_sentence_1",
"Class (film)_sentence_0"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 4437,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.628",
"value": "?search="
},
{
"operation": "search",
"time": "10.649",
"value": "Virginia Madsen"
},
{
"operation": "Now on",
"time": "13.5",
"value": "Virginia Madsen"
},
{
"operation": "Highlighting",
"time": "16.729",
"value": "Virginia Madsen_sentence_0"
},
{
"operation": "Highlighting",
"time": "20.558",
"value": "Virginia Madsen_sentence_1"
},
{
"operation": "hyperlink",
"time": "23.535",
"value": "Class"
},
{
"operation": "Now on",
"time": "24.903",
"value": "Class (film)"
},
{
"operation": "Highlighting",
"time": "27.812",
"value": "Class (film)_sentence_0"
},
{
"operation": "finish",
"time": "44.986",
"value": "finish"
}
]
}
|
exfever_4638
|
Shantel VanSanten is an Italian actress who played the role of Julie Swagger in Shooter, which aired on the American basic cable channel owned by NBCUniversal Television and Streaming division of Comcast's NBCUniversal through NBCUniversal Cable Entertainment.
|
{
"evidence": [
"Shantel VanSanten",
"USA Network"
],
"label": "REFUTE"
}
|
exfever
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nShantel VanSanten is an Italian actress who played the role of Julie Swagger in Shooter, which aired on the American basic cable channel owned by NBCUniversal Television and Streaming division of Comcast's NBCUniversal through NBCUniversal Cable Entertainment.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Shantel VanSanten",
"USA Network"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 4438,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
33001
|
More Democratic than Republican candidates won in the 1968 United States gubernatorial special elections.
|
{
"evidence": [
"By-election_sentence_0",
"1968 United States gubernatorial elections_sentence_0"
],
"label": "NOT ENOUGH INFO"
}
|
feverous
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nMore Democratic than Republican candidates won in the 1968 United States gubernatorial special elections.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"By-election_sentence_0",
"1968 United States gubernatorial elections_sentence_0"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 4439,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.632",
"value": "?search="
},
{
"operation": "search",
"time": "18.865",
"value": "Contains...1968 United States gubernatorial special elections"
},
{
"operation": "Now on",
"time": "21.568",
"value": "undefined"
},
{
"operation": "search",
"time": "46.017",
"value": "1968 gubernatorial"
},
{
"operation": "Now on",
"time": "47.23",
"value": "undefined"
},
{
"operation": "hyperlink",
"time": "50.899",
"value": "1968 United States gubernatorial elections"
},
{
"operation": "Now on",
"time": "51.604",
"value": "1968 United States gubernatorial elections"
},
{
"operation": "search",
"time": "79.935",
"value": "gubernatorial special elections"
},
{
"operation": "Now on",
"time": "80.695",
"value": "undefined"
},
{
"operation": "search",
"time": "106.448",
"value": "1968 gubernatorial special"
},
{
"operation": "Now on",
"time": "107.07",
"value": "undefined"
},
{
"operation": "search",
"time": "134.042",
"value": "Contains...1968 United States gubernatorial special elections"
},
{
"operation": "Now on",
"time": "134.878",
"value": "undefined"
},
{
"operation": "search",
"time": "178.392",
"value": "1968 United States gubernatorial elections"
},
{
"operation": "Now on",
"time": "179.019",
"value": "1968 United States gubernatorial elections"
},
{
"operation": "hyperlink",
"time": "212.087",
"value": "1968 United States elections"
},
{
"operation": "Now on",
"time": "212.841",
"value": "1968 United States elections"
},
{
"operation": "search",
"time": "237.886",
"value": "Special elections"
},
{
"operation": "Now on",
"time": "238.887",
"value": "By-election"
},
{
"operation": "Highlighting",
"time": "283.498",
"value": "By-election_sentence_0"
},
{
"operation": "search",
"time": "295.655",
"value": "1968 gubernatorial"
},
{
"operation": "Now on",
"time": "296.546",
"value": "undefined"
},
{
"operation": "search",
"time": "326.544",
"value": "1968 gubernatorial elections"
},
{
"operation": "Now on",
"time": "327.041",
"value": "undefined"
},
{
"operation": "hyperlink",
"time": "332.205",
"value": "1968 United States gubernatorial elections"
},
{
"operation": "Now on",
"time": "332.673",
"value": "1968 United States gubernatorial elections"
},
{
"operation": "Highlighting",
"time": "339.759",
"value": "1968 United States gubernatorial elections_sentence_0"
},
{
"operation": "finish",
"time": "355.509",
"value": "finish"
}
]
}
|
exfever_7866
|
Michael Edward Love co-founded the Beach Boys with his cousins Brian, Dennis, and an American musician, singer, and songwriter.
|
{
"evidence": [
"Mike Love",
"Carl Wilson"
],
"label": "SUPPORT"
}
|
exfever
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nMichael Edward Love co-founded the Beach Boys with his cousins Brian, Dennis, and an American musician, singer, and songwriter.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Mike Love",
"Carl Wilson"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 4440,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_2429
|
Harry Potter and the Deathly Hallows was released on 14 July 2017 in the United Kingdom by a British worldwide publishing house of fiction and non-fiction.
|
{
"evidence": [
"Harry Potter and the Deathly Hallows",
"Bloomsbury Publishing"
],
"label": "REFUTE"
}
|
exfever
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nHarry Potter and the Deathly Hallows was released on 14 July 2017 in the United Kingdom by a British worldwide publishing house of fiction and non-fiction.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Harry Potter and the Deathly Hallows",
"Bloomsbury Publishing"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 4441,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_4680
|
iZombie is a supernatural procedural crime drama television series developed for an American English-language commercial broadcast television network.
|
{
"evidence": [
"IZombie (TV series)",
"The CW"
],
"label": "SUPPORT"
}
|
exfever
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\niZombie is a supernatural procedural crime drama television series developed for an American English-language commercial broadcast television network.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"IZombie (TV series)",
"The CW"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 4442,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
11195
|
Dalton Kellett, a Canadian racing driver from Stouffville, Ontario, raced for Andretti Autosport for three years.
|
{
"evidence": [
"Dalton Kellett_sentence_0",
"Dalton Kellett_sentence_16"
],
"label": "SUPPORT"
}
|
feverous
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nDalton Kellett, a Canadian racing driver from Stouffville, Ontario, raced for Andretti Autosport for three years.\n",
"role": "user"
}
] |
Numerical Reasoning
|
{
"ground_truth": {
"evidence": [
"Dalton Kellett_sentence_0",
"Dalton Kellett_sentence_16"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 4443,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.722",
"value": "?search="
},
{
"operation": "search",
"time": "6.631",
"value": "Dalton Kellett,"
},
{
"operation": "Now on",
"time": "7.696",
"value": "undefined"
},
{
"operation": "hyperlink",
"time": "13.606",
"value": "Dalton Kellett"
},
{
"operation": "Now on",
"time": "14.432",
"value": "Dalton Kellett"
},
{
"operation": "Highlighting",
"time": "18.214",
"value": "Dalton Kellett_sentence_0"
},
{
"operation": "Page search",
"time": "33.291",
"value": "andretti"
},
{
"operation": "Highlighting",
"time": "52.47",
"value": "Dalton Kellett_sentence_16"
},
{
"operation": "finish",
"time": "67.424",
"value": "finish"
}
]
}
|
53230
|
Aedesius, the father of Margaret "the Virgin," disowned her after embracing Christianity, while her mother died after childbirth.
|
{
"evidence": [
"Aedesius_sentence_1",
"Aedesius of Alexandria_sentence_1",
"Aedesius of Alexandria_sentence_4",
"Aedesius of Alexandria_sentence_3"
],
"label": "NOT ENOUGH INFO"
}
|
feverous
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nAedesius, the father of Margaret \"the Virgin,\" disowned her after embracing Christianity, while her mother died after childbirth.\n",
"role": "user"
}
] |
Entity Disambiguation
|
{
"ground_truth": {
"evidence": [
"Aedesius_sentence_1",
"Aedesius of Alexandria_sentence_1",
"Aedesius of Alexandria_sentence_4",
"Aedesius of Alexandria_sentence_3"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 4444,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.996",
"value": "?search="
},
{
"operation": "search",
"time": "14.117",
"value": "Aedesius"
},
{
"operation": "Now on",
"time": "15.001",
"value": "Aedesius"
},
{
"operation": "Highlighting",
"time": "20.281",
"value": "Aedesius_sentence_1"
},
{
"operation": "search",
"time": "67.878",
"value": "Aedesius of Alexandria"
},
{
"operation": "Now on",
"time": "70.379",
"value": "Aedesius of Alexandria"
},
{
"operation": "Highlighting",
"time": "80.161",
"value": "Aedesius of Alexandria_sentence_1"
},
{
"operation": "Highlighting",
"time": "129.407",
"value": "Aedesius of Alexandria_sentence_7"
},
{
"operation": "Highlighting",
"time": "145.087",
"value": "Aedesius of Alexandria_sentence_4"
},
{
"operation": "Highlighting",
"time": "149.631",
"value": "Aedesius of Alexandria_sentence_3"
},
{
"operation": "Highlighting deleted",
"time": "153.577",
"value": "Aedesius of Alexandria_sentence_3"
},
{
"operation": "Highlighting",
"time": "155.759",
"value": "Aedesius of Alexandria_sentence_3"
},
{
"operation": "Highlighting deleted",
"time": "160.329",
"value": "Aedesius of Alexandria_sentence_7"
},
{
"operation": "finish",
"time": "176.307",
"value": "finish"
}
]
}
|
63308
|
During a performance in Newcastle upon Tyne, UK on October 31, 2007, Meatloaf Removing the jacket he was wearing, he thanked the audience for 30 years, said "goodbye forever" and left the stage
|
{
"evidence": [
"Meat Loaf_sentence_217",
"Meat Loaf_sentence_219"
],
"label": "SUPPORT"
}
|
feverous
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nDuring a performance in Newcastle upon Tyne, UK on October 31, 2007, Meatloaf Removing the jacket he was wearing, he thanked the audience for 30 years, said \"goodbye forever\" and left the stage\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Meat Loaf_sentence_217",
"Meat Loaf_sentence_219"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 4445,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.718",
"value": "?search="
},
{
"operation": "search",
"time": "18.671",
"value": "Meatloaf (singer)"
},
{
"operation": "Now on",
"time": "23.086",
"value": "Meat Loaf"
},
{
"operation": "Page search",
"time": "29.856",
"value": "newcastle"
},
{
"operation": "Highlighting",
"time": "48.513",
"value": "Meat Loaf_sentence_217"
},
{
"operation": "Highlighting",
"time": "56.113",
"value": "Meat Loaf_sentence_219"
},
{
"operation": "finish",
"time": "80.529",
"value": "finish"
}
]
}
|
exfever_6688
|
Barack Obama is an American politician born in a municipality holding secondary status in a country, state, province, department, or other sub-national entity, usually as its seat of the government in the U.S. state of Hawaii.
|
{
"evidence": [
"Barack Obama",
"Honolulu",
"Capital city"
],
"label": "REFUTE"
}
|
exfever
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nBarack Obama is an American politician born in a municipality holding secondary status in a country, state, province, department, or other sub-national entity, usually as its seat of the government in the U.S. state of Hawaii.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Barack Obama",
"Honolulu",
"Capital city"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 4446,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_7756
|
Randy Orton is an American professional wrestler and actor who is the son of an American semi-retired professional wrestler, better known by the ring name "Cowboy".
|
{
"evidence": [
"Randy Orton",
"Bob Orton Jr."
],
"label": "SUPPORT"
}
|
exfever
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nRandy Orton is an American professional wrestler and actor who is the son of an American semi-retired professional wrestler, better known by the ring name \"Cowboy\".\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Randy Orton",
"Bob Orton Jr."
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 4447,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_1829
|
Victor-Marie Hugo played Cromwell and drama Hernani at the forefront of the artistic, literary, musical, and intellectual movement that originated in Asia.
|
{
"evidence": [
"Victor Hugo",
"Romanticism"
],
"label": "REFUTE"
}
|
exfever
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nVictor-Marie Hugo played Cromwell and drama Hernani at the forefront of the artistic, literary, musical, and intellectual movement that originated in Asia.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Victor Hugo",
"Romanticism"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 4448,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_7522
|
Iain Glen is an Irish actor known for his role as an English officer in the British Indian Army who made three exploratory expeditions to Africa.
|
{
"evidence": [
"Iain Glen",
"John Hanning Speke"
],
"label": "REFUTE"
}
|
exfever
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nIain Glen is an Irish actor known for his role as an English officer in the British Indian Army who made three exploratory expeditions to Africa.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Iain Glen",
"John Hanning Speke"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 4449,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
54934
|
Sally Stanford made an arranged appearance at the men's luncheon during the 1967 California Jaycees Annual Convention which took place at the San Francisco Hilton ballroom.
|
{
"evidence": [
"Sally Stanford_sentence_10"
],
"label": "REFUTE"
}
|
feverous
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nSally Stanford made an arranged appearance at the men's luncheon during the 1967 California Jaycees Annual Convention which took place at the San Francisco Hilton ballroom.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Sally Stanford_sentence_10"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 4450,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.962",
"value": "?search="
},
{
"operation": "search",
"time": "29.213",
"value": "Sally Stanford"
},
{
"operation": "Now on",
"time": "30.366",
"value": "Sally Stanford"
},
{
"operation": "Highlighting",
"time": "48.32",
"value": "Sally Stanford_sentence_10"
},
{
"operation": "finish",
"time": "58.443",
"value": "finish"
}
]
}
|
39502
|
The Tivoli Provisioning Manager's actions are fully automated and need only minimal supervision from humans.
|
{
"evidence": [
"Tivoli Provisioning Manager_sentence_1",
"Tivoli Provisioning Manager_sentence_2",
"Hypervisor_sentence_0",
"Tivoli Provisioning Manager_sentence_10",
"Tivoli Provisioning Manager_sentence_11",
"Tivoli Provisioning Manager_sentence_15",
"Tivoli Provisioning Manager_sentence_16"
],
"label": "NOT ENOUGH INFO"
}
|
feverous
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nThe Tivoli Provisioning Manager's actions are fully automated and need only minimal supervision from humans.\n",
"role": "user"
}
] |
Multi-hop Reasoning
|
{
"ground_truth": {
"evidence": [
"Tivoli Provisioning Manager_sentence_1",
"Tivoli Provisioning Manager_sentence_2",
"Hypervisor_sentence_0",
"Tivoli Provisioning Manager_sentence_10",
"Tivoli Provisioning Manager_sentence_11",
"Tivoli Provisioning Manager_sentence_15",
"Tivoli Provisioning Manager_sentence_16"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 4451,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.624",
"value": "?search="
},
{
"operation": "search",
"time": "9.674",
"value": "Tivoli Provisioning Manager"
},
{
"operation": "Now on",
"time": "11.237",
"value": "Tivoli Provisioning Manager"
},
{
"operation": "Page search",
"time": "38.58",
"value": "auto"
},
{
"operation": "Page search",
"time": "62.853",
"value": "supervis"
},
{
"operation": "Page search",
"time": "72.628",
"value": "human"
},
{
"operation": "hyperlink",
"time": "79.454",
"value": "Tivoli Software"
},
{
"operation": "Now on",
"time": "80.599",
"value": "Tivoli Software"
},
{
"operation": "hyperlink",
"time": "101.697",
"value": "Tivoli Provisioning Manager"
},
{
"operation": "Now on",
"time": "102.975",
"value": "Tivoli Provisioning Manager"
},
{
"operation": "Highlighting",
"time": "105.265",
"value": "Tivoli Provisioning Manager_sentence_1"
},
{
"operation": "Highlighting",
"time": "109.505",
"value": "Tivoli Provisioning Manager_sentence_2"
},
{
"operation": "hyperlink",
"time": "114.652",
"value": "hypervisors"
},
{
"operation": "Now on",
"time": "116.396",
"value": "Hypervisor"
},
{
"operation": "Highlighting",
"time": "118.983",
"value": "Hypervisor_sentence_0"
},
{
"operation": "back-button-clicked",
"time": "122.388",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "122.518",
"value": "Tivoli Provisioning Manager"
},
{
"operation": "Highlighting",
"time": "143.552",
"value": "Tivoli Provisioning Manager_sentence_10"
},
{
"operation": "Highlighting",
"time": "149.138",
"value": "Tivoli Provisioning Manager_sentence_11"
},
{
"operation": "Highlighting",
"time": "160.508",
"value": "Tivoli Provisioning Manager_sentence_15"
},
{
"operation": "Highlighting",
"time": "161.565",
"value": "Tivoli Provisioning Manager_sentence_16"
},
{
"operation": "hyperlink",
"time": "170.638",
"value": "Tivoli Software"
},
{
"operation": "Now on",
"time": "171.458",
"value": "Tivoli Software"
},
{
"operation": "back-button-clicked",
"time": "185.342",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "185.479",
"value": "Tivoli Provisioning Manager"
},
{
"operation": "finish",
"time": "217.02",
"value": "finish"
}
]
}
|
20718
|
Interleukin-12 subunit alpha is a protein expressed by activated dendritic cells (DC), macrophages, and monocytes and plays an important role in sustaining IFNγ production and the generation and encoded by the IL12A gene which encodes a subunit of the cytokine Interleukin 12 (IL-12) that acts on T and natural killer cells and is required for the T-cell-dependent induction of interferon gamma .
|
{
"evidence": [
"IL12A_sentence_0",
"IL12A_sentence_1",
"IL12A_sentence_3"
],
"label": "NOT ENOUGH INFO"
}
|
feverous
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nInterleukin-12 subunit alpha is a protein expressed by activated dendritic cells (DC), macrophages, and monocytes and plays an important role in sustaining IFNγ production and the generation and encoded by the IL12A gene which encodes a subunit of the cytokine Interleukin 12 (IL-12) that acts on T and natural killer cells and is required for the T-cell-dependent induction of interferon gamma .\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"IL12A_sentence_0",
"IL12A_sentence_1",
"IL12A_sentence_3"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 4452,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.632",
"value": "?search="
},
{
"operation": "search",
"time": "6.974",
"value": "Interleukin-12 subunit alpha"
},
{
"operation": "Now on",
"time": "8.885",
"value": "undefined"
},
{
"operation": "hyperlink",
"time": "12.835",
"value": "IL12A"
},
{
"operation": "Now on",
"time": "13.502",
"value": "IL12A"
},
{
"operation": "Highlighting",
"time": "30.507",
"value": "IL12A_sentence_0"
},
{
"operation": "Highlighting",
"time": "45.042",
"value": "IL12A_sentence_1"
},
{
"operation": "Highlighting",
"time": "61.455",
"value": "IL12A_sentence_3"
},
{
"operation": "hyperlink",
"time": "80.767",
"value": "protein"
},
{
"operation": "Now on",
"time": "82.762",
"value": "Protein"
},
{
"operation": "Page search",
"time": "92.922",
"value": "dendritic"
},
{
"operation": "back-button-clicked",
"time": "115.271",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "115.373",
"value": "IL12A"
},
{
"operation": "finish",
"time": "133.731",
"value": "finish"
}
]
}
|
5444
|
Zhang Wenli was the adoptive son of Wang Rong who prevented a mutiny against the Wang clan.
|
{
"evidence": [
"Zhang Wenli_sentence_3"
],
"label": "REFUTE"
}
|
feverous
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nZhang Wenli was the adoptive son of Wang Rong who prevented a mutiny against the Wang clan.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Zhang Wenli_sentence_3"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 4453,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.511",
"value": "?search="
},
{
"operation": "search",
"time": "13.902",
"value": "Zhang Wenli"
},
{
"operation": "Now on",
"time": "14.766",
"value": "Zhang Wenli"
},
{
"operation": "Highlighting",
"time": "24.743",
"value": "Zhang Wenli_sentence_1"
},
{
"operation": "Highlighting deleted",
"time": "48.773",
"value": "Zhang Wenli_sentence_1"
},
{
"operation": "Highlighting",
"time": "49.763",
"value": "Zhang Wenli_sentence_3"
},
{
"operation": "finish",
"time": "112.351",
"value": "finish"
}
]
}
|
exfever_6501
|
Clinton Eastwood Jr. is an American actor who was successful in the Western TV series that aired for eight seasons on the CBS network on Friday nights, from January 9, 1959, to September 3, 1965.
|
{
"evidence": [
"Clint Eastwood",
"Rawhide (TV series)"
],
"label": "SUPPORT"
}
|
exfever
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nClinton Eastwood Jr. is an American actor who was successful in the Western TV series that aired for eight seasons on the CBS network on Friday nights, from January 9, 1959, to September 3, 1965.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Clint Eastwood",
"Rawhide (TV series)"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 4454,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
75083
|
The Wilmington and Conway Railroad construction was never completed due to lack of funds.
|
{
"evidence": [
"Wilmington and Conway Railroad_sentence_1",
"Wilmington and Conway Railroad_sentence_0"
],
"label": "NOT ENOUGH INFO"
}
|
feverous
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nThe Wilmington and Conway Railroad construction was never completed due to lack of funds.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Wilmington and Conway Railroad_sentence_1",
"Wilmington and Conway Railroad_sentence_0"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 4455,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "3.321",
"value": "undefined"
},
{
"operation": "search",
"time": "36.367",
"value": "The Wilmington and Conway Railroad"
},
{
"operation": "Now on",
"time": "40.392",
"value": "undefined"
},
{
"operation": "hyperlink",
"time": "42.87",
"value": "Wilmington and Conway Railroad"
},
{
"operation": "Now on",
"time": "43.834",
"value": "Wilmington and Conway Railroad"
},
{
"operation": "Highlighting",
"time": "62.799",
"value": "Wilmington and Conway Railroad_sentence_1"
},
{
"operation": "Highlighting",
"time": "70.182",
"value": "Wilmington and Conway Railroad_sentence_0"
},
{
"operation": "finish",
"time": "90.109",
"value": "finish"
}
]
}
|
exfever_936
|
Shia LaBeouf is a French actor who played Louis Stevens in an American comedy television series that aired on Disney Channel.
|
{
"evidence": [
"Shia LaBeouf",
"Even Stevens"
],
"label": "REFUTE"
}
|
exfever
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nShia LaBeouf is a French actor who played Louis Stevens in an American comedy television series that aired on Disney Channel.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Shia LaBeouf",
"Even Stevens"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 4456,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_5989
|
Alan Rickman was an English actor and director who became a member of a major British theatre company, based in Stratford-upon-Avon, Warwickshire, England.
|
{
"evidence": [
"Alan Rickman",
"Royal Shakespeare Company"
],
"label": "SUPPORT"
}
|
exfever
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nAlan Rickman was an English actor and director who became a member of a major British theatre company, based in Stratford-upon-Avon, Warwickshire, England.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Alan Rickman",
"Royal Shakespeare Company"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 4457,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_6026
|
Slash is best known as the lead guitarist of an American hard rock band that was disbanded in 1985.
|
{
"evidence": [
"Slash (musician)",
"Guns N' Roses"
],
"label": "REFUTE"
}
|
exfever
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nSlash is best known as the lead guitarist of an American hard rock band that was disbanded in 1985.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Slash (musician)",
"Guns N' Roses"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 4458,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_5659
|
Nick Kroll is an American actor who is also known as the voice actor in a 2016 children computer-animated comedy film co-written by a Canadian screenwriter who collaborated with his childhood friend Seth Rogen for various films.
|
{
"evidence": [
"Nick Kroll",
"Sausage Party",
"Evan Goldberg"
],
"label": "REFUTE"
}
|
exfever
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nNick Kroll is an American actor who is also known as the voice actor in a 2016 children computer-animated comedy film co-written by a Canadian screenwriter who collaborated with his childhood friend Seth Rogen for various films.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Nick Kroll",
"Sausage Party",
"Evan Goldberg"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 4459,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_6358
|
The film Just My Luck stars Lindsay Lohan as Ashley Albright, the luckiest person in the most densely populated and geographically smallest of the five boroughs of Albany.
|
{
"evidence": [
"Just My Luck (2006 film)",
"Manhattan"
],
"label": "REFUTE"
}
|
exfever
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nThe film Just My Luck stars Lindsay Lohan as Ashley Albright, the luckiest person in the most densely populated and geographically smallest of the five boroughs of Albany.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Just My Luck (2006 film)",
"Manhattan"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 4460,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
87637
|
ADCYAP1R1, also known as PAC1, is a protein (a large biomolecule or macromolecule that consists of one or more long chains of amino acid residues) that shares traints significant homology with members of the G-protein coupled class B glocagon/secretin receptor family, and is expressed in the adrenal medulla, pancreatic acini, uterus, myenteric plexus, and brain.
|
{
"evidence": [
"ADCYAP1R1_sentence_0",
"ADCYAP1R1_sentence_2",
"ADCYAP1R1_sentence_5",
"Protein_sentence_3"
],
"label": "SUPPORT"
}
|
feverous
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nADCYAP1R1, also known as PAC1, is a protein (a large biomolecule or macromolecule that consists of one or more long chains of amino acid residues) that shares traints significant homology with members of the G-protein coupled class B glocagon/secretin receptor family, and is expressed in the adrenal medulla, pancreatic acini, uterus, myenteric plexus, and brain.\n",
"role": "user"
}
] |
Multi-hop Reasoning
|
{
"ground_truth": {
"evidence": [
"ADCYAP1R1_sentence_0",
"ADCYAP1R1_sentence_2",
"ADCYAP1R1_sentence_5",
"Protein_sentence_3"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 4461,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.917",
"value": "?search="
},
{
"operation": "search",
"time": "40.612",
"value": "ADCYAP1R1"
},
{
"operation": "Now on",
"time": "41.227",
"value": "ADCYAP1R1"
},
{
"operation": "Highlighting",
"time": "47.98",
"value": "ADCYAP1R1_sentence_0"
},
{
"operation": "Highlighting",
"time": "71.83",
"value": "ADCYAP1R1_sentence_2"
},
{
"operation": "Page search",
"time": "116.349",
"value": "adrenal"
},
{
"operation": "Highlighting",
"time": "118.483",
"value": "ADCYAP1R1_sentence_5"
},
{
"operation": "hyperlink",
"time": "128.285",
"value": "protein"
},
{
"operation": "Now on",
"time": "129.935",
"value": "Protein"
},
{
"operation": "Highlighting",
"time": "132.757",
"value": "Protein_sentence_3"
},
{
"operation": "finish",
"time": "142.687",
"value": "finish"
}
]
}
|
exfever_6162
|
Quebec is one of the thirteen provinces and territories of Australia where much of the population lives in urban areas along the St. Lawrence River, between Quebec City and the second-most populous city in Canada, and the most populous city in the Canadian province of Quebec.
|
{
"evidence": [
"Quebec",
"Montreal"
],
"label": "REFUTE"
}
|
exfever
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nQuebec is one of the thirteen provinces and territories of Australia where much of the population lives in urban areas along the St. Lawrence River, between Quebec City and the second-most populous city in Canada, and the most populous city in the Canadian province of Quebec.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Quebec",
"Montreal"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 4462,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_8228
|
Nigeria is a country in Antarctica where the capital and eighth-most populous city of Nigeria is based.
|
{
"evidence": [
"Nigeria",
"Abuja"
],
"label": "REFUTE"
}
|
exfever
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nNigeria is a country in Antarctica where the capital and eighth-most populous city of Nigeria is based.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Nigeria",
"Abuja"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 4463,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
42006
|
Joe Motzko played for two other teams before joining the Washington Capitals on July 9, 2007.
|
{
"evidence": [
"Joe Motzko_sentence_11",
"Joe Motzko_sentence_6",
"Joe Motzko_sentence_9"
],
"label": "SUPPORT"
}
|
feverous
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nJoe Motzko played for two other teams before joining the Washington Capitals on July 9, 2007.\n",
"role": "user"
}
] |
Numerical Reasoning
|
{
"ground_truth": {
"evidence": [
"Joe Motzko_sentence_11",
"Joe Motzko_sentence_6",
"Joe Motzko_sentence_9"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 4464,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.892",
"value": "?search="
},
{
"operation": "search",
"time": "7.532",
"value": "Joe Motzko"
},
{
"operation": "Now on",
"time": "8.4",
"value": "Joe Motzko"
},
{
"operation": "Highlighting",
"time": "22.322",
"value": "Joe Motzko_sentence_11"
},
{
"operation": "Highlighting",
"time": "36.405",
"value": "Joe Motzko_sentence_4"
},
{
"operation": "Highlighting",
"time": "44.841",
"value": "Joe Motzko_sentence_6"
},
{
"operation": "Highlighting deleted",
"time": "46.011",
"value": "Joe Motzko_sentence_4"
},
{
"operation": "Highlighting",
"time": "59.015",
"value": "Joe Motzko_sentence_9"
},
{
"operation": "finish",
"time": "69.099",
"value": "finish"
}
]
}
|
17925
|
Cordia Harrington established the Bun Lady Scholarship, and was awarded the Beta Gamma Sigma (an academic honor society for top business school students founded at the University of Wisconsin in 1913) Medallion for Entrepreneurship in 2013.
|
{
"evidence": [
"Cordia Harrington_sentence_5"
],
"label": "REFUTE"
}
|
feverous
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nCordia Harrington established the Bun Lady Scholarship, and was awarded the Beta Gamma Sigma (an academic honor society for top business school students founded at the University of Wisconsin in 1913) Medallion for Entrepreneurship in 2013.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Cordia Harrington_sentence_5"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 4465,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.61",
"value": "?search="
},
{
"operation": "search",
"time": "22.999",
"value": "Cordia Harrington"
},
{
"operation": "Now on",
"time": "24.314",
"value": "Cordia Harrington"
},
{
"operation": "Page search",
"time": "35.415",
"value": "bun"
},
{
"operation": "Highlighting",
"time": "42.224",
"value": "Cordia Harrington_sentence_6"
},
{
"operation": "Highlighting",
"time": "52.724",
"value": "Cordia Harrington_sentence_5"
},
{
"operation": "hyperlink",
"time": "60.374",
"value": "Beta Gamma Sigma"
},
{
"operation": "Now on",
"time": "61.41",
"value": "Beta Gamma Sigma"
},
{
"operation": "Highlighting",
"time": "67.295",
"value": "Beta Gamma Sigma_sentence_0"
},
{
"operation": "Highlighting",
"time": "67.296",
"value": "Beta Gamma Sigma_title"
},
{
"operation": "Highlighting",
"time": "73.561",
"value": "Beta Gamma Sigma_sentence_1"
},
{
"operation": "back-button-clicked",
"time": "89.596",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "89.639",
"value": "Cordia Harrington"
},
{
"operation": "Highlighting",
"time": "100.794",
"value": "Cordia Harrington_sentence_5"
},
{
"operation": "finish",
"time": "111.548",
"value": "finish"
}
]
}
|
exfever_5692
|
Harley Quinn is a comic relief henchwoman for the superhero Joker in an American animated television series whose genre of speculative fiction examines the adventures, personalities, and ethics of costumed crime fighters known as superheroes.
|
{
"evidence": [
"Harley Quinn",
"Batman: The Animated Series",
"Superhero fiction"
],
"label": "REFUTE"
}
|
exfever
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nHarley Quinn is a comic relief henchwoman for the superhero Joker in an American animated television series whose genre of speculative fiction examines the adventures, personalities, and ethics of costumed crime fighters known as superheroes.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Harley Quinn",
"Batman: The Animated Series",
"Superhero fiction"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 4466,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_7313
|
Edgar Allan Poe is regarded in the US as a central figure in the artistic, literary, musical, and intellectual movements that originated in Europe towards the end of the 17th century.
|
{
"evidence": [
"Edgar Allan Poe",
"Romanticism"
],
"label": "REFUTE"
}
|
exfever
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nEdgar Allan Poe is regarded in the US as a central figure in the artistic, literary, musical, and intellectual movements that originated in Europe towards the end of the 17th century.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Edgar Allan Poe",
"Romanticism"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 4467,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_5852
|
Naresh Malhotra’s first film directed was Yeh Dillagi.
|
{
"evidence": [
"Yeh Dillagi",
"Sabrina (1954 film)"
],
"label": "NOT ENOUGH INFO"
}
|
exfever
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nNaresh Malhotra’s first film directed was Yeh Dillagi.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Yeh Dillagi",
"Sabrina (1954 film)"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 4468,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
29002
|
Dwight–Hooker Avenue Historic District is a national historic district located in New York, and includes 17 contributing residential buildings most of which were built between 1895 and 1815.
|
{
"evidence": [
"Dwight–Hooker Avenue Historic District_sentence_2"
],
"label": "REFUTE"
}
|
feverous
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nDwight–Hooker Avenue Historic District is a national historic district located in New York, and includes 17 contributing residential buildings most of which were built between 1895 and 1815.\n",
"role": "user"
}
] |
Numerical Reasoning
|
{
"ground_truth": {
"evidence": [
"Dwight–Hooker Avenue Historic District_sentence_2"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 4469,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.885",
"value": "?search="
},
{
"operation": "Now on",
"time": "68.759",
"value": "?search="
},
{
"operation": "search",
"time": "80.588",
"value": "Dwight–Hooker Avenue Historic District"
},
{
"operation": "Now on",
"time": "81.212",
"value": "Dwight–Hooker Avenue Historic District"
},
{
"operation": "Highlighting",
"time": "128.641",
"value": "Dwight–Hooker Avenue Historic District_sentence_2"
},
{
"operation": "finish",
"time": "136.192",
"value": "finish"
}
]
}
|
exfever_1121
|
Leroy Jethro Gibbs is a fictional character who is a former U.S. Marine Corps Scout Sniper turned special agent who commands a team for the primary law enforcement agency of the U.S. Department of the Navy.?
|
{
"evidence": [
"Leroy Jethro Gibbs",
"Naval Criminal Investigative Service"
],
"label": "SUPPORT"
}
|
exfever
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nLeroy Jethro Gibbs is a fictional character who is a former U.S. Marine Corps Scout Sniper turned special agent who commands a team for the primary law enforcement agency of the U.S. Department of the Navy.?\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Leroy Jethro Gibbs",
"Naval Criminal Investigative Service"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 4470,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_1453
|
Matt Damon's first film was The Martian.
|
{
"evidence": [
"The Martian (film)",
"Drew Goddard"
],
"label": "NOT ENOUGH INFO"
}
|
exfever
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nMatt Damon's first film was The Martian.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"The Martian (film)",
"Drew Goddard"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 4471,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
14003
|
HTR3A causes fast, depolarizing responses in neurons after activation and the A subunit forms homomers with a very high single channel conductance of 0.6pS.
|
{
"evidence": [
"HTR3A_sentence_4"
],
"label": "REFUTE"
}
|
feverous
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nHTR3A causes fast, depolarizing responses in neurons after activation and the A subunit forms homomers with a very high single channel conductance of 0.6pS.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"HTR3A_sentence_4"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 4472,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.078",
"value": "?search="
},
{
"operation": "search",
"time": "13.825",
"value": "HTR3A"
},
{
"operation": "Now on",
"time": "15.124",
"value": "HTR3A"
},
{
"operation": "Page search",
"time": "29.622",
"value": "depolarizing"
},
{
"operation": "Highlighting",
"time": "34.992",
"value": "HTR3A_sentence_3"
},
{
"operation": "Highlighting",
"time": "40.391",
"value": "HTR3A_sentence_4"
},
{
"operation": "Highlighting deleted",
"time": "54.604",
"value": "HTR3A_sentence_3"
},
{
"operation": "finish",
"time": "68.859",
"value": "finish"
}
]
}
|
exfever_2099
|
Jeffrey Lynn Goldblum is an American actor who has starred in several films, including a 2016 American science fiction action film directed by Will Smith.
|
{
"evidence": [
"Jeff Goldblum",
"Independence Day: Resurgence"
],
"label": "REFUTE"
}
|
exfever
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nJeffrey Lynn Goldblum is an American actor who has starred in several films, including a 2016 American science fiction action film directed by Will Smith.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Jeff Goldblum",
"Independence Day: Resurgence"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 4473,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
1724
|
Dachi of Iberia took a more conciliatory line with his Iranian suzerains and also spent special missionary efforts to further Christianize his mountainous subjects.
|
{
"evidence": [
"Dachi of Iberia_sentence_3",
"Dachi of Iberia_sentence_4"
],
"label": "SUPPORT"
}
|
feverous
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nDachi of Iberia took a more conciliatory line with his Iranian suzerains and also spent special missionary efforts to further Christianize his mountainous subjects.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Dachi of Iberia_sentence_3",
"Dachi of Iberia_sentence_4"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 4474,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "search",
"time": "7.285",
"value": "Dachi of Iberia"
},
{
"operation": "Page search",
"time": "21.35",
"value": "suzerains"
},
{
"operation": "Highlighting",
"time": "28.325",
"value": "Dachi of Iberia_sentence_3"
},
{
"operation": "Highlighting",
"time": "43.282",
"value": "Dachi of Iberia_sentence_4"
},
{
"operation": "finish",
"time": "49.08",
"value": "finish"
}
]
}
|
exfever_4863
|
Judy Greer will also feature in a sports drama.
|
{
"evidence": [
"Judy Greer",
"Adaptation (film)",
"Comedy-drama"
],
"label": "NOT ENOUGH INFO"
}
|
exfever
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nJudy Greer will also feature in a sports drama.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Judy Greer",
"Adaptation (film)",
"Comedy-drama"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 4475,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_6924
|
Lil Wayne's career began in 1994 when he joined an American record label founded in 1991 by brothers Slim and Birdman Williams.
|
{
"evidence": [
"Lil Wayne",
"Cash Money Records"
],
"label": "REFUTE"
}
|
exfever
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nLil Wayne's career began in 1994 when he joined an American record label founded in 1991 by brothers Slim and Birdman Williams.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Lil Wayne",
"Cash Money Records"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 4476,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_3335
|
Matthew Modine starred as Louden Swain in a 1955 romantic drama film that also starred an American actor known for his performance as Drew Ballinger in Deliverance.
|
{
"evidence": [
"Matthew Modine",
"Vision Quest",
"Ronny Cox"
],
"label": "REFUTE"
}
|
exfever
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nMatthew Modine starred as Louden Swain in a 1955 romantic drama film that also starred an American actor known for his performance as Drew Ballinger in Deliverance.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Matthew Modine",
"Vision Quest",
"Ronny Cox"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 4477,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
19944
|
Cullingworth is a village and civil parish in West Yorkshire, England.
|
{
"evidence": [
"Cullingworth_sentence_0"
],
"label": "SUPPORT"
}
|
feverous
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nCullingworth is a village and civil parish in West Yorkshire, England.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Cullingworth_sentence_0"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 4478,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.077",
"value": "?search="
},
{
"operation": "search",
"time": "8.754",
"value": "Cullingworth"
},
{
"operation": "Now on",
"time": "9.657",
"value": "Cullingworth"
},
{
"operation": "Highlighting",
"time": "11.202",
"value": "Cullingworth_sentence_0"
},
{
"operation": "finish",
"time": "23.375",
"value": "finish"
}
]
}
|
11528
|
The IWRG 5th Anniversary Show featured six matches that involved 28 wrestlers.
|
{
"evidence": [
"IWRG 5th Anniversary Show_sentence_16"
],
"label": "REFUTE"
}
|
feverous
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nThe IWRG 5th Anniversary Show featured six matches that involved 28 wrestlers.\n",
"role": "user"
}
] |
Numerical Reasoning
|
{
"ground_truth": {
"evidence": [
"IWRG 5th Anniversary Show_sentence_16"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 4479,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.7",
"value": "?search="
},
{
"operation": "search",
"time": "13.813",
"value": "IWRG 5th Anniversary Show"
},
{
"operation": "Now on",
"time": "15.131",
"value": "IWRG 5th Anniversary Show"
},
{
"operation": "hyperlink",
"time": "20.913",
"value": "4 Results IWRG 5th Anniversary Show_section_5"
},
{
"operation": "hyperlink",
"time": "85.438",
"value": "2 Event IWRG 5th Anniversary Show_section_3"
},
{
"operation": "Highlighting",
"time": "98.59",
"value": "IWRG 5th Anniversary Show_header_cell_1_1_0"
},
{
"operation": "Highlighting",
"time": "98.592",
"value": "IWRG 5th Anniversary Show_section_5"
},
{
"operation": "Highlighting",
"time": "100.201",
"value": "IWRG 5th Anniversary Show_header_cell_1_2_0"
},
{
"operation": "Highlighting",
"time": "100.726",
"value": "IWRG 5th Anniversary Show_header_cell_1_3_0"
},
{
"operation": "Highlighting",
"time": "101.212",
"value": "IWRG 5th Anniversary Show_header_cell_1_4_0"
},
{
"operation": "Highlighting",
"time": "101.761",
"value": "IWRG 5th Anniversary Show_header_cell_1_5_0"
},
{
"operation": "Highlighting deleted",
"time": "104.958",
"value": "IWRG 5th Anniversary Show_header_cell_1_1_0"
},
{
"operation": "Highlighting deleted",
"time": "105.364",
"value": "IWRG 5th Anniversary Show_header_cell_1_2_0"
},
{
"operation": "Highlighting deleted",
"time": "105.734",
"value": "IWRG 5th Anniversary Show_header_cell_1_3_0"
},
{
"operation": "Highlighting deleted",
"time": "106.141",
"value": "IWRG 5th Anniversary Show_header_cell_1_4_0"
},
{
"operation": "Highlighting deleted",
"time": "106.566",
"value": "IWRG 5th Anniversary Show_header_cell_1_5_0"
},
{
"operation": "Highlighting deleted",
"time": "106.572",
"value": "IWRG 5th Anniversary Show_section_5"
},
{
"operation": "Highlighting",
"time": "126.521",
"value": "IWRG 5th Anniversary Show_sentence_3"
},
{
"operation": "Highlighting",
"time": "130.218",
"value": "IWRG 5th Anniversary Show_sentence_4"
},
{
"operation": "Highlighting",
"time": "162.203",
"value": "IWRG 5th Anniversary Show_sentence_16"
},
{
"operation": "Highlighting deleted",
"time": "235.855",
"value": "IWRG 5th Anniversary Show_sentence_4"
},
{
"operation": "Highlighting deleted",
"time": "236.425",
"value": "IWRG 5th Anniversary Show_sentence_3"
},
{
"operation": "finish",
"time": "247.312",
"value": "finish"
}
]
}
|
exfever_931
|
Vienna is the capital of a country in the southern part of Central Europe and also the center of its culture that has been influenced by its past and present neighbor countries.
|
{
"evidence": [
"Vienna",
"Culture of Austria",
"Austria"
],
"label": "SUPPORT"
}
|
exfever
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nVienna is the capital of a country in the southern part of Central Europe and also the center of its culture that has been influenced by its past and present neighbor countries.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Vienna",
"Culture of Austria",
"Austria"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 4480,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_7914
|
George Harrison was an English musician who achieved international fame as the lead guitarist of the English rock band formed in 1960 in a city and metropolitan borough in Merseyside, England.
|
{
"evidence": [
"George Harrison",
"The Beatles",
"Liverpool"
],
"label": "SUPPORT"
}
|
exfever
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nGeorge Harrison was an English musician who achieved international fame as the lead guitarist of the English rock band formed in 1960 in a city and metropolitan borough in Merseyside, England.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"George Harrison",
"The Beatles",
"Liverpool"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 4481,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
44115
|
Oklahoma State Highway 54 runs north to south and exists in two parts which lie at approximately the same longitude which is a geographic coordinate that specifies the east-west position of a point on the Earth's surface or the surface of a celestial body.
|
{
"evidence": [
"Oklahoma State Highway 54_sentence_1",
"Longitude_sentence_3"
],
"label": "SUPPORT"
}
|
feverous
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nOklahoma State Highway 54 runs north to south and exists in two parts which lie at approximately the same longitude which is a geographic coordinate that specifies the east-west position of a point on the Earth's surface or the surface of a celestial body.\n",
"role": "user"
}
] |
Multi-hop Reasoning
|
{
"ground_truth": {
"evidence": [
"Oklahoma State Highway 54_sentence_1",
"Longitude_sentence_3"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 4482,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.889",
"value": "?search="
},
{
"operation": "search",
"time": "7.527",
"value": "Oklahoma State Highway 54"
},
{
"operation": "Now on",
"time": "8.754",
"value": "Oklahoma State Highway 54"
},
{
"operation": "Highlighting",
"time": "12.135",
"value": "Oklahoma State Highway 54_sentence_1"
},
{
"operation": "hyperlink",
"time": "19.778",
"value": "longitude"
},
{
"operation": "Now on",
"time": "21.593",
"value": "Longitude"
},
{
"operation": "Highlighting",
"time": "23.713",
"value": "Longitude_sentence_3"
},
{
"operation": "finish",
"time": "29.68",
"value": "finish"
}
]
}
|
35727
|
Švenčionys Ghetto had 566 Jews - some of which established contracts with active Lithuanian policemen, who frustrated economical exploitation plans of the Germans towards Soviet territories, while the others joined Jewish partisans.
|
{
"evidence": [
"Švenčionys Ghetto_sentence_14",
"Švenčionys Ghetto_sentence_16",
"Soviet partisans_sentence_3"
],
"label": "NOT ENOUGH INFO"
}
|
feverous
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nŠvenčionys Ghetto had 566 Jews - some of which established contracts with active Lithuanian policemen, who frustrated economical exploitation plans of the Germans towards Soviet territories, while the others joined Jewish partisans.\n",
"role": "user"
}
] |
Multi-hop Reasoning
|
{
"ground_truth": {
"evidence": [
"Švenčionys Ghetto_sentence_14",
"Švenčionys Ghetto_sentence_16",
"Soviet partisans_sentence_3"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 4483,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.643",
"value": "?search="
},
{
"operation": "search",
"time": "13.53",
"value": "Svencionys Ghetto"
},
{
"operation": "Now on",
"time": "14.443",
"value": "Švenčionys Ghetto"
},
{
"operation": "Highlighting",
"time": "34.154",
"value": "Švenčionys Ghetto_sentence_14"
},
{
"operation": "Highlighting",
"time": "135.623",
"value": "Švenčionys Ghetto_sentence_16"
},
{
"operation": "hyperlink",
"time": "138.411",
"value": "Soviet partisans"
},
{
"operation": "Now on",
"time": "139.197",
"value": "Soviet partisans"
},
{
"operation": "Highlighting",
"time": "188.613",
"value": "Soviet partisans_sentence_3"
},
{
"operation": "finish",
"time": "262.06",
"value": "finish"
}
]
}
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.