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 |
|---|---|---|---|---|---|---|---|---|
exfever_4755
|
Andrew Ayers Stanton's film work includes directing a 2016 computer-animated adventure film released by an American film production studio and subsidiary of Walt Disney Studios, which is owned by The Warner Bros. Company.
|
{
"evidence": [
"Andrew Stanton",
"Finding Dory",
"Walt Disney Pictures"
],
"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:\nAndrew Ayers Stanton's film work includes directing a 2016 computer-animated adventure film released by an American film production studio and subsidiary of Walt Disney Studios, which is owned by The Warner Bros. Company.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Andrew Stanton",
"Finding Dory",
"Walt Disney Pictures"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 3484,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_3676
|
Lost Highway is a 1997 neo-noir film directed by David Lynch, financed by a French film production and distribution company founded in 1993 by Francis Bouygues.
|
{
"evidence": [
"Lost Highway (film)",
"Ciby 2000"
],
"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:\nLost Highway is a 1997 neo-noir film directed by David Lynch, financed by a French film production and distribution company founded in 1993 by Francis Bouygues.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Lost Highway (film)",
"Ciby 2000"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 3485,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_3851
|
Beautiful" is fourth studio album of a Canadian singer, songwriter, actress, and television personality.
|
{
"evidence": [
"Beautiful (Christina Aguilera song)",
"Christina Aguilera"
],
"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:\nBeautiful\" is fourth studio album of a Canadian singer, songwriter, actress, and television personality.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Beautiful (Christina Aguilera song)",
"Christina Aguilera"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 3486,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
26206
|
Jean Tarcisius Batamboc is a Cameroonian sprinter who competed at the 2017 World Championships, held at the London Olympic Stadium on 4−5 August, ending last in the heats.
|
{
"evidence": [
"Jean Tarcisius Batamboc_sentence_0",
"Jean Tarcisius Batamboc_sentence_2",
"2017 World Championships in Athletics – Men's 100 metres_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:\nJean Tarcisius Batamboc is a Cameroonian sprinter who competed at the 2017 World Championships, held at the London Olympic Stadium on 4−5 August, ending last in the heats.\n",
"role": "user"
}
] |
Multi-hop Reasoning
|
{
"ground_truth": {
"evidence": [
"Jean Tarcisius Batamboc_sentence_0",
"Jean Tarcisius Batamboc_sentence_2",
"2017 World Championships in Athletics – Men's 100 metres_sentence_0"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 3487,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.627",
"value": "?search="
},
{
"operation": "search",
"time": "18.286",
"value": "Jean Tarcisius Batamboc"
},
{
"operation": "Now on",
"time": "19.205",
"value": "Jean Tarcisius Batamboc"
},
{
"operation": "Highlighting",
"time": "26.984",
"value": "Jean Tarcisius Batamboc_sentence_0"
},
{
"operation": "Highlighting",
"time": "32.731",
"value": "Jean Tarcisius Batamboc_sentence_2"
},
{
"operation": "hyperlink",
"time": "39.168",
"value": "2017 World Championships"
},
{
"operation": "Now on",
"time": "40.433",
"value": "2017 World Championships in Athletics – Men's 100 metres"
},
{
"operation": "Highlighting",
"time": "42.652",
"value": "2017 World Championships in Athletics – Men's 100 metres_sentence_0"
},
{
"operation": "finish",
"time": "50.88",
"value": "finish"
}
]
}
|
exfever_7990
|
The film Seven Days in May is about a cabal's planned takeover of the United States government in reaction to the president's negotiation with a socialist and communist state.
|
{
"evidence": [
"Seven Days in May",
"Soviet Union"
],
"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 film Seven Days in May is about a cabal's planned takeover of the United States government in reaction to the president's negotiation with a socialist and communist state.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Seven Days in May",
"Soviet Union"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 3488,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
4254
|
Newsjack was first broadcast on BBC Radio 7 (a British digital radio station) and had an "Closed door" policy on writing and was designed to give new writers an opportunity
|
{
"evidence": [
"Newsjack_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:\nNewsjack was first broadcast on BBC Radio 7 (a British digital radio station) and had an \"Closed door\" policy on writing and was designed to give new writers an opportunity\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Newsjack_sentence_5"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 3489,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "search",
"time": "18.138",
"value": "Newsjack"
},
{
"operation": "Highlighting",
"time": "27.294",
"value": "Newsjack_sentence_4"
},
{
"operation": "hyperlink",
"time": "29.617",
"value": "BBC Radio 7"
},
{
"operation": "Highlighting",
"time": "32.617",
"value": "BBC Radio 4 Extra_sentence_2"
},
{
"operation": "Highlighting",
"time": "32.635",
"value": "BBC Radio 4 Extra_title"
},
{
"operation": "back-button-clicked",
"time": "34.059",
"value": "back-button-clicked"
},
{
"operation": "Highlighting",
"time": "37.501",
"value": "Newsjack_sentence_5"
},
{
"operation": "finish",
"time": "56.633",
"value": "finish"
}
]
}
|
26174
|
The elevated train station, Nantou, was the main gate of Zhonsghan when it started operations in 2011.
|
{
"evidence": [
"Nantou railway station_sentence_0",
"Nantou railway station_sentence_3",
"Nantou railway station_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 elevated train station, Nantou, was the main gate of Zhonsghan when it started operations in 2011.\n",
"role": "user"
}
] |
Search terms not in claim
|
{
"ground_truth": {
"evidence": [
"Nantou railway station_sentence_0",
"Nantou railway station_sentence_3",
"Nantou railway station_sentence_4"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 3490,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.649",
"value": "?search="
},
{
"operation": "search",
"time": "6.249",
"value": "Nantou"
},
{
"operation": "Now on",
"time": "6.856",
"value": "Nantou"
},
{
"operation": "search",
"time": "25.297",
"value": "Nantou railway station"
},
{
"operation": "Now on",
"time": "25.94",
"value": "Nantou railway station"
},
{
"operation": "Highlighting",
"time": "32.883",
"value": "Nantou railway station_sentence_0"
},
{
"operation": "Highlighting",
"time": "42.897",
"value": "Nantou railway station_sentence_3"
},
{
"operation": "Highlighting",
"time": "46.499",
"value": "Nantou railway station_sentence_4"
},
{
"operation": "finish",
"time": "51.102",
"value": "finish"
}
]
}
|
exfever_8582
|
Bohemian Rhapsody is a movie consisting of a segment of verse, often a narrative set to music.
|
{
"evidence": [
"Bohemian Rhapsody",
"Ballad"
],
"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:\nBohemian Rhapsody is a movie consisting of a segment of verse, often a narrative set to music.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Bohemian Rhapsody",
"Ballad"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 3491,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
80830
|
Gisela of Swabia was crowned thirteen days after Conrad's coronation by Archbishop Pilgrim of Cologne, resulting in Henry II replacing her as her son's regent with Poppo of Trier, who was consecrated by Erkanbald, Archbishop of Mainz.
|
{
"evidence": [
"Gisela of Swabia_sentence_11",
"Gisela of Swabia_sentence_12",
"Poppo (archbishop of Trier)_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:\nGisela of Swabia was crowned thirteen days after Conrad's coronation by Archbishop Pilgrim of Cologne, resulting in Henry II replacing her as her son's regent with Poppo of Trier, who was consecrated by Erkanbald, Archbishop of Mainz.\n",
"role": "user"
}
] |
Multi-hop Reasoning
|
{
"ground_truth": {
"evidence": [
"Gisela of Swabia_sentence_11",
"Gisela of Swabia_sentence_12",
"Poppo (archbishop of Trier)_sentence_4"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 3492,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.076",
"value": "?search="
},
{
"operation": "search",
"time": "9.355",
"value": "Gisela of Swabia"
},
{
"operation": "Now on",
"time": "11.117",
"value": "Gisela of Swabia"
},
{
"operation": "Highlighting",
"time": "31.566",
"value": "Gisela of Swabia_sentence_11"
},
{
"operation": "Highlighting",
"time": "85.776",
"value": "Gisela of Swabia_sentence_12"
},
{
"operation": "hyperlink",
"time": "156.563",
"value": "Poppo of Trier"
},
{
"operation": "Now on",
"time": "157.316",
"value": "Poppo (archbishop of Trier)"
},
{
"operation": "Highlighting",
"time": "160.31",
"value": "Poppo (archbishop of Trier)_sentence_4"
},
{
"operation": "finish",
"time": "171.379",
"value": "finish"
}
]
}
|
exfever_2882
|
Jonah Hill is considered by critics as one of the best comedians in the world.
|
{
"evidence": [
"21 Jump Street (film)",
"Patrick Hasburgh"
],
"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:\nJonah Hill is considered by critics as one of the best comedians in the world.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"21 Jump Street (film)",
"Patrick Hasburgh"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 3493,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
95151
|
In 1957 Christina Stone (who was married to Loke Wan Tho, a Singaporean business magnate, and later to Jeffrey Stone a film director, then her final marriage was to the Parsi businessman) made a presentation at the Odeon Theatre in Singapore to the winners of a rock'n'roll competition, and in 1965 she appeared in the Vogue Magazine as one of Singapore's most beautiful woman.
|
{
"evidence": [
"Christina Stone_sentence_0",
"Christina Stone_sentence_1",
"Christina Stone_sentence_13",
"Christina Stone_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:\nIn 1957 Christina Stone (who was married to Loke Wan Tho, a Singaporean business magnate, and later to Jeffrey Stone a film director, then her final marriage was to the Parsi businessman) made a presentation at the Odeon Theatre in Singapore to the winners of a rock'n'roll competition, and in 1965 she appeared in the Vogue Magazine as one of Singapore's most beautiful woman.\n",
"role": "user"
}
] |
Multi-hop Reasoning
|
{
"ground_truth": {
"evidence": [
"Christina Stone_sentence_0",
"Christina Stone_sentence_1",
"Christina Stone_sentence_13",
"Christina Stone_sentence_16"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 3494,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.379",
"value": "?search="
},
{
"operation": "search",
"time": "24.572",
"value": "Christina Stone"
},
{
"operation": "Now on",
"time": "25.83",
"value": "Christina Stone"
},
{
"operation": "Page search",
"time": "34.207",
"value": "loke"
},
{
"operation": "Highlighting",
"time": "38.896",
"value": "Christina Stone_sentence_0"
},
{
"operation": "Highlighting",
"time": "56.431",
"value": "Christina Stone_sentence_1"
},
{
"operation": "Page search",
"time": "66.939",
"value": "odeon"
},
{
"operation": "Highlighting",
"time": "72.08",
"value": "Christina Stone_sentence_13"
},
{
"operation": "Highlighting",
"time": "83.711",
"value": "Christina Stone_sentence_16"
},
{
"operation": "finish",
"time": "198.843",
"value": "finish"
}
]
}
|
78372
|
The 1964 Tower Hamlets Council election took place on May 7, 1964 and was the first to the newly formed borough, where 115 candidates stood in the election for 60 seats being contested.
|
{
"evidence": [
"1964 Tower Hamlets London Borough Council election_sentence_0",
"1964 Tower Hamlets London Borough Council election_sentence_2",
"1964 Tower Hamlets London Borough Council election_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:\nThe 1964 Tower Hamlets Council election took place on May 7, 1964 and was the first to the newly formed borough, where 115 candidates stood in the election for 60 seats being contested.\n",
"role": "user"
}
] |
Multi-hop Reasoning
|
{
"ground_truth": {
"evidence": [
"1964 Tower Hamlets London Borough Council election_sentence_0",
"1964 Tower Hamlets London Borough Council election_sentence_2",
"1964 Tower Hamlets London Borough Council election_sentence_5"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 3495,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.584",
"value": "?search="
},
{
"operation": "search",
"time": "35.857",
"value": "1964 Tower Hamlets Council election"
},
{
"operation": "Now on",
"time": "37.191",
"value": "1964 Tower Hamlets London Borough Council election"
},
{
"operation": "Highlighting",
"time": "40.023",
"value": "1964 Tower Hamlets London Borough Council election_sentence_0"
},
{
"operation": "Highlighting",
"time": "51.348",
"value": "1964 Tower Hamlets London Borough Council election_sentence_2"
},
{
"operation": "Highlighting",
"time": "58.255",
"value": "1964 Tower Hamlets London Borough Council election_sentence_5"
},
{
"operation": "finish",
"time": "66.355",
"value": "finish"
}
]
}
|
7874
|
Julia burni belongs to the Juliidae family of the gastropoda class, a class which comprises snails and slugs from saltwater, and from freshwater only.
|
{
"evidence": [
"Gastropoda_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:\nJulia burni belongs to the Juliidae family of the gastropoda class, a class which comprises snails and slugs from saltwater, and from freshwater only.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Gastropoda_sentence_1"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 3496,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.552",
"value": "?search="
},
{
"operation": "search",
"time": "5.359",
"value": "Julia burni"
},
{
"operation": "Now on",
"time": "6.047",
"value": "Julia burni"
},
{
"operation": "Highlighting",
"time": "11.733",
"value": "Julia burni_cell_0_7_0"
},
{
"operation": "Highlighting",
"time": "11.734",
"value": "Julia burni_title"
},
{
"operation": "Highlighting",
"time": "11.736",
"value": "Julia burni_header_cell_0_1_0"
},
{
"operation": "Highlighting",
"time": "11.736",
"value": "Julia burni_header_cell_0_0_0"
},
{
"operation": "Highlighting",
"time": "12.516",
"value": "Julia burni_cell_0_7_1"
},
{
"operation": "Highlighting",
"time": "16.079",
"value": "Julia burni_cell_0_4_0"
},
{
"operation": "Highlighting",
"time": "16.666",
"value": "Julia burni_cell_0_4_1"
},
{
"operation": "hyperlink",
"time": "21.223",
"value": "Gastropoda"
},
{
"operation": "Now on",
"time": "22.056",
"value": "Gastropoda"
},
{
"operation": "Highlighting",
"time": "23.954",
"value": "Gastropoda_sentence_0"
},
{
"operation": "Highlighting deleted",
"time": "30.668",
"value": "Gastropoda_sentence_0"
},
{
"operation": "Highlighting",
"time": "31.396",
"value": "Gastropoda_sentence_1"
},
{
"operation": "finish",
"time": "200.659",
"value": "finish"
}
]
}
|
exfever_1998
|
A Thousand Suns is the fourth studio album by the American rock band Linkin Park produced by an American record producer who is also the former co-president of an American record label owned by Sony Music Entertainment, a subsidiary of Sony Corporation of America, the North American division of Japanese conglomerate Sony.
|
{
"evidence": [
"A Thousand Suns",
"Rick Rubin",
"Columbia Records"
],
"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:\nA Thousand Suns is the fourth studio album by the American rock band Linkin Park produced by an American record producer who is also the former co-president of an American record label owned by Sony Music Entertainment, a subsidiary of Sony Corporation of America, the North American division of Japanese conglomerate Sony.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"A Thousand Suns",
"Rick Rubin",
"Columbia Records"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 3497,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_1468
|
Li Lianjie is a Chinese film actor who is a naturalized citizen in a sovereign island city-state in maritime Southeast Asia that has four official languages, including a language that is classified into two distinct dialects.
|
{
"evidence": [
"Jet Li",
"Singapore",
"Singaporean Mandarin"
],
"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:\nLi Lianjie is a Chinese film actor who is a naturalized citizen in a sovereign island city-state in maritime Southeast Asia that has four official languages, including a language that is classified into two distinct dialects.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Jet Li",
"Singapore",
"Singaporean Mandarin"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 3498,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_3396
|
A Liberty Bond was a war bond that was sold in the United States to support the Allied cause in a war that began on the 28th of July, 1914, which was the deadliest conflict in history, with an estimated 9 million killed in combat.
|
{
"evidence": [
"Liberty bond",
"World War I"
],
"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:\nA Liberty Bond was a war bond that was sold in the United States to support the Allied cause in a war that began on the 28th of July, 1914, which was the deadliest conflict in history, with an estimated 9 million killed in combat.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Liberty bond",
"World War I"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 3499,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_5315
|
The Prince of Egypt is a 1998 American animated musical drama film that follows the life of the considered most important prophet in Judaism.
|
{
"evidence": [
"The Prince of Egypt",
"Moses"
],
"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 Prince of Egypt is a 1998 American animated musical drama film that follows the life of the considered most important prophet in Judaism.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"The Prince of Egypt",
"Moses"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 3500,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_5225
|
The film, The Hateful Eight, stars an American actor who began acting in an American Western television series based on a Pulitzer Prize-winning novel of the same name.
|
{
"evidence": [
"The Hateful Eight",
"Kurt Russell",
"The Travels of Jaimie McPheeters (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:\nThe film, The Hateful Eight, stars an American actor who began acting in an American Western television series based on a Pulitzer Prize-winning novel of the same name.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"The Hateful Eight",
"Kurt Russell",
"The Travels of Jaimie McPheeters (TV series)"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 3501,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
16323
|
Hereford Castle is a manor rebuilt by William FitzOsbern, Earl of Hereford and father of Roger FitzOsbern who was involved in an unsuccessful attempt to depose King William .
|
{
"evidence": [
"Hereford Castle_sentence_12",
"Hereford Castle_sentence_13",
"Hereford Castle_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:\nHereford Castle is a manor rebuilt by William FitzOsbern, Earl of Hereford and father of Roger FitzOsbern who was involved in an unsuccessful attempt to depose King William .\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Hereford Castle_sentence_12",
"Hereford Castle_sentence_13",
"Hereford Castle_sentence_14"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 3502,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.604",
"value": "?search="
},
{
"operation": "Now on",
"time": "10.859",
"value": "?search="
},
{
"operation": "search",
"time": "31.123",
"value": "Hereford Castle"
},
{
"operation": "Now on",
"time": "32.077",
"value": "Hereford Castle"
},
{
"operation": "Page search",
"time": "46.961",
"value": "William FitzOsbern"
},
{
"operation": "Highlighting",
"time": "71.808",
"value": "Hereford Castle_sentence_12"
},
{
"operation": "Highlighting",
"time": "72.742",
"value": "Hereford Castle_sentence_13"
},
{
"operation": "Highlighting",
"time": "73.669",
"value": "Hereford Castle_sentence_14"
},
{
"operation": "finish",
"time": "144.352",
"value": "finish"
}
]
}
|
exfever_7031
|
Lily Collins is an actress born in the town of Surrey in the southwest of London on the A3 trunk road between the capital of Portsmouth.
|
{
"evidence": [
"Lily Collins",
"Guildford"
],
"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:\nLily Collins is an actress born in the town of Surrey in the southwest of London on the A3 trunk road between the capital of Portsmouth.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Lily Collins",
"Guildford"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 3503,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
67697
|
Marie Amelie is a Russian-born writer and blogger who lived as an illegal immigrant in Norway between 2002 and 2011,She was deported from Norway to Russia on 24 January 2011.
|
{
"evidence": [
"Maria Amelie_sentence_0",
"Maria Amelie_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:\nMarie Amelie is a Russian-born writer and blogger who lived as an illegal immigrant in Norway between 2002 and 2011,She was deported from Norway to Russia on 24 January 2011.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Maria Amelie_sentence_0",
"Maria Amelie_sentence_1"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 3504,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.221",
"value": "?search="
},
{
"operation": "search",
"time": "13.058",
"value": "Marie Amelie"
},
{
"operation": "Now on",
"time": "13.997",
"value": "Maria Amalia"
},
{
"operation": "search",
"time": "101.213",
"value": "Marie Amelie is a Russian-born writer and blogger"
},
{
"operation": "Now on",
"time": "103.175",
"value": "undefined"
},
{
"operation": "search",
"time": "138.789",
"value": "Maria Amelie"
},
{
"operation": "Now on",
"time": "139.759",
"value": "Maria Amelie"
},
{
"operation": "Highlighting",
"time": "268.239",
"value": "Maria Amelie_sentence_0"
},
{
"operation": "Highlighting",
"time": "296.824",
"value": "Maria Amelie_sentence_1"
},
{
"operation": "finish",
"time": "323.564",
"value": "finish"
}
]
}
|
exfever_2688
|
Cape Fear is the first American noir psychological thriller film released in 1962.
|
{
"evidence": [
"Cape Fear (1991 film)",
"Cape Fear (1962 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:\nCape Fear is the first American noir psychological thriller film released in 1962.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Cape Fear (1991 film)",
"Cape Fear (1962 film)"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 3505,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_1607
|
Trolls is a 2016 American computer-animated musical comedy film directed by an American film director who directed Shrek Forever After.
|
{
"evidence": [
"Trolls (film)",
"Mike Mitchell (director)"
],
"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:\nTrolls is a 2016 American computer-animated musical comedy film directed by an American film director who directed Shrek Forever After.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Trolls (film)",
"Mike Mitchell (director)"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 3506,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_7981
|
Jordan rose to fame using a vocal register occupying the frequency range just below the modal voice register.
|
{
"evidence": [
"Jordan Knight",
"Falsetto"
],
"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:\nJordan rose to fame using a vocal register occupying the frequency range just below the modal voice register.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Jordan Knight",
"Falsetto"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 3507,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_7820
|
Frank Ocean also won the Soul Train Music Awards.
|
{
"evidence": [
"Frank Ocean",
"Def Jam Recordings"
],
"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:\nFrank Ocean also won the Soul Train Music Awards.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Frank Ocean",
"Def Jam Recordings"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 3508,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
17754
|
For the draws, clubs were separated into seeded and unseeded teams based on their club coefficient after the two separate qualifying tournaments in the new system, which is known as the Champions Path and the Non-Champions Path
|
{
"evidence": [
"2009–10 UEFA Champions League_sentence_18",
"2009–10 UEFA Champions League_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:\nFor the draws, clubs were separated into seeded and unseeded teams based on their club coefficient after the two separate qualifying tournaments in the new system, which is known as the Champions Path and the Non-Champions Path\n",
"role": "user"
}
] |
Search terms not in claim
|
{
"ground_truth": {
"evidence": [
"2009–10 UEFA Champions League_sentence_18",
"2009–10 UEFA Champions League_sentence_15"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 3509,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.412",
"value": "?search="
},
{
"operation": "search",
"time": "28.914",
"value": "Contains...Champions Path"
},
{
"operation": "Now on",
"time": "30.505",
"value": "undefined"
},
{
"operation": "search",
"time": "90.307",
"value": "Contains...Champions Path and the Non-Champions Path"
},
{
"operation": "Now on",
"time": "92.304",
"value": "undefined"
},
{
"operation": "search",
"time": "132.773",
"value": "2017–18 UEFA Champions League qualifying phase and play-off round"
},
{
"operation": "Now on",
"time": "135.178",
"value": "2017–18 UEFA Champions League qualifying phase and play-off round"
},
{
"operation": "Page search",
"time": "143.954",
"value": "For the draws, clubs were separated into seeded and unseeded teams"
},
{
"operation": "Page search",
"time": "144.54",
"value": "For the draws, clubs were separated into seeded and unseeded teams"
},
{
"operation": "search",
"time": "170.281",
"value": "For the draws, clubs were separated into seeded and unseeded teams"
},
{
"operation": "Now on",
"time": "172.187",
"value": "undefined"
},
{
"operation": "hyperlink",
"time": "175.685",
"value": "Teams 2009–10 UEFA Champions League_section_3"
},
{
"operation": "Now on",
"time": "178.89",
"value": "2009–10 UEFA Champions League#Teams 2009.E2.80.9310 UEFA Champions League section 3"
},
{
"operation": "Page search",
"time": "208.9",
"value": "seeded"
},
{
"operation": "Highlighting",
"time": "217.123",
"value": "2009–10 UEFA Champions League_sentence_18"
},
{
"operation": "Highlighting",
"time": "240.691",
"value": "2009–10 UEFA Champions League_sentence_15"
},
{
"operation": "Page search",
"time": "302.8",
"value": "Champions Path"
},
{
"operation": "Highlighting deleted",
"time": "321.363",
"value": "2009–10 UEFA Champions League_sentence_15"
},
{
"operation": "Highlighting",
"time": "322.482",
"value": "2009–10 UEFA Champions League_sentence_15"
},
{
"operation": "finish",
"time": "332.191",
"value": "finish"
}
]
}
|
49184
|
Victor Hanzeli never joined the University of Washington faculty.
|
{
"evidence": [
"Victor Hanzeli_sentence_0",
"Victor Hanzeli_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:\nVictor Hanzeli never joined the University of Washington faculty.\n",
"role": "user"
}
] |
Entity Disambiguation
|
{
"ground_truth": {
"evidence": [
"Victor Hanzeli_sentence_0",
"Victor Hanzeli_sentence_8"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 3510,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.882",
"value": "?search="
},
{
"operation": "search",
"time": "12.218",
"value": "Victor Hanzeli"
},
{
"operation": "Now on",
"time": "13.401",
"value": "Victor Hanzeli"
},
{
"operation": "Highlighting",
"time": "19.689",
"value": "Victor Hanzeli_sentence_0"
},
{
"operation": "Highlighting",
"time": "28.252",
"value": "Victor Hanzeli_sentence_8"
},
{
"operation": "finish",
"time": "45.7",
"value": "finish"
}
]
}
|
87926
|
Junior Eurovision Song Contest 2011 was held in the previous year's winning country and Public Television of Armenia ARMTV was the main organizer and Bulgaria (first appearance 2007) returned.
|
{
"evidence": [
"Junior Eurovision Song Contest 2011_sentence_1",
"Junior Eurovision Song Contest 2011_sentence_2",
"Junior Eurovision Song Contest 2011_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:\nJunior Eurovision Song Contest 2011 was held in the previous year's winning country and Public Television of Armenia ARMTV was the main organizer and Bulgaria (first appearance 2007) returned.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Junior Eurovision Song Contest 2011_sentence_1",
"Junior Eurovision Song Contest 2011_sentence_2",
"Junior Eurovision Song Contest 2011_sentence_3"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 3511,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.072",
"value": "?search="
},
{
"operation": "search",
"time": "5.969",
"value": "Junior Eurovision Song Contest 2011"
},
{
"operation": "Now on",
"time": "6.782",
"value": "Junior Eurovision Song Contest 2011"
},
{
"operation": "Highlighting",
"time": "19.053",
"value": "Junior Eurovision Song Contest 2011_sentence_1"
},
{
"operation": "Highlighting",
"time": "26.183",
"value": "Junior Eurovision Song Contest 2011_sentence_2"
},
{
"operation": "Highlighting",
"time": "172.808",
"value": "Junior Eurovision Song Contest 2011_sentence_3"
},
{
"operation": "finish",
"time": "199.658",
"value": "finish"
}
]
}
|
74966
|
One Direction is an English-Irish boy band formed in 2010 who won the NME Award for Worst Band and whose single, "What Makes You Beautiful" ranked 1st in the UK and Irish chart in 2011
|
{
"evidence": [
"One Direction_sentence_1",
"What Makes You Beautiful_sentence_68",
"What Makes You Beautiful_sentence_72"
],
"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:\nOne Direction is an English-Irish boy band formed in 2010 who won the NME Award for Worst Band and whose single, \"What Makes You Beautiful\" ranked 1st in the UK and Irish chart in 2011\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"One Direction_sentence_1",
"What Makes You Beautiful_sentence_68",
"What Makes You Beautiful_sentence_72"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 3512,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.966",
"value": "?search="
},
{
"operation": "search",
"time": "16.162",
"value": "One Direction"
},
{
"operation": "Now on",
"time": "18.318",
"value": "One Direction"
},
{
"operation": "Highlighting",
"time": "23.069",
"value": "One Direction_sentence_1"
},
{
"operation": "Highlighting deleted",
"time": "43.819",
"value": "One Direction_sentence_1"
},
{
"operation": "Page search",
"time": "47.877",
"value": "nme"
},
{
"operation": "Page search",
"time": "122.826",
"value": "what makes you"
},
{
"operation": "Page search",
"time": "224.435",
"value": "new musical"
},
{
"operation": "Highlighting",
"time": "280.657",
"value": "One Direction_sentence_1"
},
{
"operation": "Highlighting",
"time": "288.384",
"value": "One Direction_sentence_4"
},
{
"operation": "Highlighting deleted",
"time": "312.639",
"value": "One Direction_sentence_4"
},
{
"operation": "Page search",
"time": "334.585",
"value": "what makes"
},
{
"operation": "search",
"time": "370.914",
"value": "What Makes You Beautiful"
},
{
"operation": "Now on",
"time": "371.956",
"value": "What Makes You Beautiful"
},
{
"operation": "Highlighting",
"time": "398.221",
"value": "What Makes You Beautiful_sentence_68"
},
{
"operation": "Highlighting",
"time": "406.192",
"value": "What Makes You Beautiful_sentence_72"
},
{
"operation": "search",
"time": "430.042",
"value": "NME Awards one direction"
},
{
"operation": "Now on",
"time": "432.45",
"value": "undefined"
},
{
"operation": "search",
"time": "451.4",
"value": "NME Awards"
},
{
"operation": "Now on",
"time": "452.796",
"value": "NME Awards"
},
{
"operation": "search",
"time": "500.575",
"value": "nme award 2013"
},
{
"operation": "Now on",
"time": "501.73",
"value": "undefined"
},
{
"operation": "search",
"time": "515.663",
"value": "one direction nme"
},
{
"operation": "Now on",
"time": "516.21",
"value": "undefined"
},
{
"operation": "hyperlink",
"time": "528.055",
"value": "Massive Attack"
},
{
"operation": "Now on",
"time": "529.191",
"value": "Massive Attack"
},
{
"operation": "finish",
"time": "546.19",
"value": "finish"
}
]
}
|
83900
|
Leophron is the eldest of the two sons of Anaxilas, in whose name Micythus assumed the sovereignty, and who afterward, at the instigation of Hieron of Syracuse, dispossessed the latter of his authority, Diodorus, from whom we learn these facts, does not mention the name of either of the young princes.
|
{
"evidence": [
"Leophron_sentence_2"
],
"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:\nLeophron is the eldest of the two sons of Anaxilas, in whose name Micythus assumed the sovereignty, and who afterward, at the instigation of Hieron of Syracuse, dispossessed the latter of his authority, Diodorus, from whom we learn these facts, does not mention the name of either of the young princes.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Leophron_sentence_2"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 3513,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.929",
"value": "?search="
},
{
"operation": "search",
"time": "14.475",
"value": "Leophron"
},
{
"operation": "Now on",
"time": "15.632",
"value": "Leophron"
},
{
"operation": "Highlighting",
"time": "32.469",
"value": "Leophron_sentence_2"
},
{
"operation": "finish",
"time": "213.966",
"value": "finish"
}
]
}
|
68089
|
O Amor Está no Ar is a Brazilian Telenovela (serial drama or soap opera) released on September 6, 1997 created by Alcides Nogueira and directed by Wolf Maya.
|
{
"evidence": [
"O Amor Está no Ar_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:\nO Amor Está no Ar is a Brazilian Telenovela (serial drama or soap opera) released on September 6, 1997 created by Alcides Nogueira and directed by Wolf Maya.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"O Amor Está no Ar_sentence_0"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 3514,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.74",
"value": "?search="
},
{
"operation": "search",
"time": "12.856",
"value": "O Amor Está no Ar"
},
{
"operation": "Now on",
"time": "14.296",
"value": "O Amor Está no Ar"
},
{
"operation": "Highlighting",
"time": "17.085",
"value": "O Amor Está no Ar_sentence_0"
},
{
"operation": "Highlighting",
"time": "36.099",
"value": "O Amor Está no Ar_cell_0_2_1"
},
{
"operation": "Highlighting",
"time": "36.139",
"value": "O Amor Está no Ar_header_cell_0_2_0"
},
{
"operation": "Highlighting",
"time": "36.139",
"value": "O Amor Está no Ar_header_cell_0_0_0"
},
{
"operation": "Highlighting",
"time": "52.909",
"value": "O Amor Está no Ar_cell_0_3_1"
},
{
"operation": "Highlighting",
"time": "52.948",
"value": "O Amor Está no Ar_header_cell_0_3_0"
},
{
"operation": "Highlighting",
"time": "61.317",
"value": "O Amor Está no Ar_cell_0_14_1"
},
{
"operation": "Highlighting",
"time": "61.354",
"value": "O Amor Está no Ar_header_cell_0_14_0"
},
{
"operation": "Highlighting",
"time": "61.356",
"value": "O Amor Está no Ar_header_cell_0_12_0"
},
{
"operation": "Highlighting",
"time": "80.06",
"value": "O Amor Está no Ar_cell_0_7_1"
},
{
"operation": "Highlighting",
"time": "80.106",
"value": "O Amor Está no Ar_header_cell_0_7_0"
},
{
"operation": "Highlighting deleted",
"time": "83.366",
"value": "O Amor Está no Ar_sentence_0"
},
{
"operation": "Highlighting",
"time": "84.684",
"value": "O Amor Está no Ar_cell_0_1_1"
},
{
"operation": "Highlighting",
"time": "84.734",
"value": "O Amor Está no Ar_header_cell_0_1_0"
},
{
"operation": "Highlighting deleted",
"time": "152.767",
"value": "O Amor Está no Ar_cell_0_1_1"
},
{
"operation": "Highlighting deleted",
"time": "152.807",
"value": "O Amor Está no Ar_header_cell_0_1_0"
},
{
"operation": "Highlighting deleted",
"time": "153.398",
"value": "O Amor Está no Ar_cell_0_2_1"
},
{
"operation": "Highlighting deleted",
"time": "153.439",
"value": "O Amor Está no Ar_header_cell_0_2_0"
},
{
"operation": "Highlighting deleted",
"time": "153.83",
"value": "O Amor Está no Ar_cell_0_3_1"
},
{
"operation": "Highlighting deleted",
"time": "153.875",
"value": "O Amor Está no Ar_header_cell_0_3_0"
},
{
"operation": "Highlighting deleted",
"time": "155.953",
"value": "O Amor Está no Ar_cell_0_7_1"
},
{
"operation": "Highlighting deleted",
"time": "155.993",
"value": "O Amor Está no Ar_header_cell_0_7_0"
},
{
"operation": "Highlighting deleted",
"time": "155.997",
"value": "O Amor Está no Ar_header_cell_0_0_0"
},
{
"operation": "Highlighting deleted",
"time": "157.792",
"value": "O Amor Está no Ar_cell_0_14_1"
},
{
"operation": "Highlighting deleted",
"time": "157.844",
"value": "O Amor Está no Ar_header_cell_0_14_0"
},
{
"operation": "Highlighting deleted",
"time": "157.85",
"value": "O Amor Está no Ar_header_cell_0_12_0"
},
{
"operation": "Highlighting",
"time": "161.813",
"value": "O Amor Está no Ar_sentence_0"
},
{
"operation": "finish",
"time": "178.321",
"value": "finish"
}
]
}
|
2100
|
Heinrich Ehrlich (a piano teacher at the Stern Conservatory, an private painting school) settled in Berlin in 1862 and acquired a reputation as a piano player.
|
{
"evidence": [
"Stern Conservatory_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:\nHeinrich Ehrlich (a piano teacher at the Stern Conservatory, an private painting school) settled in Berlin in 1862 and acquired a reputation as a piano player.\n",
"role": "user"
}
] |
Multi-hop Reasoning
|
{
"ground_truth": {
"evidence": [
"Stern Conservatory_sentence_0"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 3515,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "search",
"time": "13.697",
"value": "Heinrich Ehrlich"
},
{
"operation": "Highlighting",
"time": "39.543",
"value": "Heinrich Ehrlich_sentence_7"
},
{
"operation": "Highlighting",
"time": "39.544",
"value": "Heinrich Ehrlich_title"
},
{
"operation": "Highlighting",
"time": "39.545",
"value": "Heinrich Ehrlich_section_0"
},
{
"operation": "hyperlink",
"time": "45.112",
"value": "Stern Conservatory"
},
{
"operation": "Highlighting",
"time": "56.807",
"value": "Stern Conservatory_sentence_0"
},
{
"operation": "finish",
"time": "87.141",
"value": "finish"
}
]
}
|
36607
|
In "Saved by the Great White Hope," Ben and Lily use the milk from coconuts as a substitute for blood, since the milk has the same electrolyte balance as blood.
|
{
"evidence": [
"Saved by the Great White Hope_sentence_22",
"Saved by the Great White Hope_sentence_23"
],
"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:\nIn \"Saved by the Great White Hope,\" Ben and Lily use the milk from coconuts as a substitute for blood, since the milk has the same electrolyte balance as blood.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Saved by the Great White Hope_sentence_22",
"Saved by the Great White Hope_sentence_23"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 3516,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "2.842",
"value": "?search="
},
{
"operation": "search",
"time": "13.732",
"value": "Saved by the Great White Hope"
},
{
"operation": "Now on",
"time": "15.709",
"value": "Saved by the Great White Hope"
},
{
"operation": "Highlighting",
"time": "40.976",
"value": "Saved by the Great White Hope_sentence_22"
},
{
"operation": "Highlighting",
"time": "47.541",
"value": "Saved by the Great White Hope_sentence_23"
},
{
"operation": "finish",
"time": "63.589",
"value": "finish"
}
]
}
|
19668
|
The sovereign country Australia, the world's sixth largest country by area, only exported gold to Europe previously but the gold deposits were exhausted in the 1800s.
|
{
"evidence": [
"Australian gold rushes_sentence_267"
],
"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 sovereign country Australia, the world's sixth largest country by area, only exported gold to Europe previously but the gold deposits were exhausted in the 1800s.\n",
"role": "user"
}
] |
Multi-hop Reasoning
|
{
"ground_truth": {
"evidence": [
"Australian gold rushes_sentence_267"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 3517,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.347",
"value": "?search="
},
{
"operation": "search",
"time": "18.705",
"value": "Australia"
},
{
"operation": "Now on",
"time": "22.372",
"value": "Australia"
},
{
"operation": "Highlighting",
"time": "32.128",
"value": "Australia_sentence_3"
},
{
"operation": "Highlighting",
"time": "32.134",
"value": "Australia_title"
},
{
"operation": "hyperlink",
"time": "99.381",
"value": "gold rush"
},
{
"operation": "Now on",
"time": "103.639",
"value": "Australian gold rushes"
},
{
"operation": "back-button-clicked",
"time": "254.364",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "254.823",
"value": "Australia"
},
{
"operation": "Highlighting",
"time": "276.861",
"value": "Australia_sentence_78"
},
{
"operation": "Highlighting",
"time": "276.864",
"value": "Australia_section_1"
},
{
"operation": "Highlighting",
"time": "276.864",
"value": "Australia_section_4"
},
{
"operation": "Highlighting",
"time": "280.293",
"value": "Australia_sentence_79"
},
{
"operation": "hyperlink",
"time": "308.28",
"value": "series of gold rushes"
},
{
"operation": "Now on",
"time": "309.542",
"value": "Australian gold rushes"
},
{
"operation": "Highlighting",
"time": "344.157",
"value": "Australian gold rushes_sentence_264"
},
{
"operation": "Highlighting",
"time": "349.196",
"value": "Australian gold rushes_sentence_266"
},
{
"operation": "Highlighting",
"time": "353.313",
"value": "Australian gold rushes_sentence_267"
},
{
"operation": "Highlighting deleted",
"time": "355.27",
"value": "Australian gold rushes_sentence_266"
},
{
"operation": "Highlighting deleted",
"time": "357.104",
"value": "Australian gold rushes_sentence_264"
},
{
"operation": "finish",
"time": "392.99",
"value": "finish"
}
]
}
|
exfever_5926
|
X-Men is a 2000 American superhero film directed by Bryan Singer that features an ensemble cast, including the Australian actor who has headlined films in various genres, including the 2004 action gothic horror film written and directed by Bruce Davison.
|
{
"evidence": [
"X-Men (film)",
"Hugh Jackman",
"Van Helsing (film)"
],
"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:\nX-Men is a 2000 American superhero film directed by Bryan Singer that features an ensemble cast, including the Australian actor who has headlined films in various genres, including the 2004 action gothic horror film written and directed by Bruce Davison.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"X-Men (film)",
"Hugh Jackman",
"Van Helsing (film)"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 3518,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
43969
|
Sebastian Coe has fewer gold medals for the 1500 metres at the Olympics than Kipchoge Keino and Fermin Cacho.
|
{
"evidence": [
"Sebastian Coe_sentence_1",
"Kipchoge Keino_sentence_16",
"Fermín Cacho_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:\nSebastian Coe has fewer gold medals for the 1500 metres at the Olympics than Kipchoge Keino and Fermin Cacho.\n",
"role": "user"
}
] |
Numerical Reasoning
|
{
"ground_truth": {
"evidence": [
"Sebastian Coe_sentence_1",
"Kipchoge Keino_sentence_16",
"Fermín Cacho_sentence_1"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 3519,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.234",
"value": "?search="
},
{
"operation": "search",
"time": "9.25",
"value": "Sebastian Coe"
},
{
"operation": "Now on",
"time": "10.235",
"value": "Sebastian Coe"
},
{
"operation": "Highlighting",
"time": "327.86",
"value": "Sebastian Coe_sentence_1"
},
{
"operation": "search",
"time": "337.3",
"value": "Kipchoge Keino"
},
{
"operation": "Now on",
"time": "338.095",
"value": "Kipchoge Keino"
},
{
"operation": "Highlighting",
"time": "350.783",
"value": "Kipchoge Keino_sentence_12"
},
{
"operation": "Highlighting deleted",
"time": "362.791",
"value": "Kipchoge Keino_sentence_12"
},
{
"operation": "Highlighting",
"time": "372.416",
"value": "Kipchoge Keino_sentence_16"
},
{
"operation": "search",
"time": "388.226",
"value": "Fermín Cacho"
},
{
"operation": "Now on",
"time": "389.129",
"value": "Fermín Cacho"
},
{
"operation": "Highlighting",
"time": "392.905",
"value": "Fermín Cacho_sentence_1"
},
{
"operation": "finish",
"time": "450.064",
"value": "finish"
}
]
}
|
65368
|
Prador Moon is a romantic novel by Neal Asher.
|
{
"evidence": [
"Prador Moon_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:\nPrador Moon is a romantic novel by Neal Asher.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Prador Moon_sentence_0"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 3520,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.335",
"value": "?search="
},
{
"operation": "search",
"time": "11.432",
"value": "Prador Moon"
},
{
"operation": "Now on",
"time": "12.483",
"value": "Prador Moon"
},
{
"operation": "Highlighting",
"time": "16.781",
"value": "Prador Moon_sentence_0"
},
{
"operation": "finish",
"time": "23.865",
"value": "finish"
}
]
}
|
49763
|
Joe DeSantis was the fifth head coach of the Quinnipiac University men's basketball team and was named NEC Coach of the Year by CBSSportsLine.com in 1999 to 2000.
|
{
"evidence": [
"Joe DeSantis_sentence_4",
"Joe DeSantis_sentence_7"
],
"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 DeSantis was the fifth head coach of the Quinnipiac University men's basketball team and was named NEC Coach of the Year by CBSSportsLine.com in 1999 to 2000.\n",
"role": "user"
}
] |
Entity Disambiguation
|
{
"ground_truth": {
"evidence": [
"Joe DeSantis_sentence_4",
"Joe DeSantis_sentence_7"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 3521,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.756",
"value": "?search="
},
{
"operation": "search",
"time": "7.576",
"value": "Joe De Santis"
},
{
"operation": "Now on",
"time": "9.615",
"value": "Joe De Santis"
},
{
"operation": "hyperlink",
"time": "34.26",
"value": "Joe DeSantis"
},
{
"operation": "Now on",
"time": "34.898",
"value": "Joe DeSantis"
},
{
"operation": "Highlighting",
"time": "43.85",
"value": "Joe DeSantis_sentence_4"
},
{
"operation": "Highlighting",
"time": "51.264",
"value": "Joe DeSantis_sentence_7"
},
{
"operation": "finish",
"time": "57.363",
"value": "finish"
}
]
}
|
41162
|
On January 12, 1997, Jason Dawe scored just ten seconds into the overtime period to give the Sabres a 3–2 road win over the Phoenix Coyotes, and the Sabres allowed the fewest shorthanded goals (4) and scored the most shorthanded goals (16) in the NHL.
|
{
"evidence": [
"1996–97 Buffalo Sabres season_sentence_6",
"1996–97 Buffalo Sabres season_sentence_8"
],
"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:\nOn January 12, 1997, Jason Dawe scored just ten seconds into the overtime period to give the Sabres a 3–2 road win over the Phoenix Coyotes, and the Sabres allowed the fewest shorthanded goals (4) and scored the most shorthanded goals (16) in the NHL.\n",
"role": "user"
}
] |
Entity Disambiguation
|
{
"ground_truth": {
"evidence": [
"1996–97 Buffalo Sabres season_sentence_6",
"1996–97 Buffalo Sabres season_sentence_8"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 3522,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.638",
"value": "?search="
},
{
"operation": "search",
"time": "19.814",
"value": "Jason Dawe (ice hockey)"
},
{
"operation": "Now on",
"time": "20.896",
"value": "Jason Dawe (ice hockey)"
},
{
"operation": "Page search",
"time": "32.715",
"value": "1997"
},
{
"operation": "hyperlink",
"time": "46.512",
"value": "Buffalo Sabres"
},
{
"operation": "Now on",
"time": "47.55",
"value": "Buffalo Sabres"
},
{
"operation": "hyperlink",
"time": "66.242",
"value": "List of Buffalo Sabres seasons"
},
{
"operation": "Now on",
"time": "67.241",
"value": "List of Buffalo Sabres seasons"
},
{
"operation": "hyperlink",
"time": "82.36",
"value": "1996–97"
},
{
"operation": "Now on",
"time": "83.396",
"value": "1996–97 NHL season"
},
{
"operation": "Page search",
"time": "87.015",
"value": "dawe"
},
{
"operation": "back-button-clicked",
"time": "95.386",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "95.461",
"value": "List of Buffalo Sabres seasons"
},
{
"operation": "hyperlink",
"time": "97.94",
"value": "1997–98"
},
{
"operation": "Now on",
"time": "98.982",
"value": "1997–98 NHL season"
},
{
"operation": "Page search",
"time": "105.303",
"value": "dawe"
},
{
"operation": "Page search",
"time": "129.063",
"value": "jason dawe"
},
{
"operation": "search",
"time": "132.38",
"value": "jason dawe"
},
{
"operation": "Now on",
"time": "133.129",
"value": "Jason Dawe"
},
{
"operation": "hyperlink",
"time": "135.303",
"value": "Jason Dawe (ice hockey)"
},
{
"operation": "Now on",
"time": "135.659",
"value": "Jason Dawe (ice hockey)"
},
{
"operation": "Page search",
"time": "142.531",
"value": "overtime"
},
{
"operation": "Highlighting",
"time": "146.764",
"value": "Jason Dawe (ice hockey)_sentence_0"
},
{
"operation": "Highlighting",
"time": "146.765",
"value": "Jason Dawe (ice hockey)_title"
},
{
"operation": "search",
"time": "183.404",
"value": "1996–97 Buffalo Sabres season"
},
{
"operation": "Now on",
"time": "185.088",
"value": "1996–97 Buffalo Sabres season"
},
{
"operation": "Page search",
"time": "189.247",
"value": "dawe"
},
{
"operation": "Highlighting",
"time": "193.417",
"value": "1996–97 Buffalo Sabres season_sentence_6"
},
{
"operation": "Highlighting",
"time": "211.229",
"value": "1996–97 Buffalo Sabres season_sentence_8"
},
{
"operation": "finish",
"time": "221.508",
"value": "finish"
}
]
}
|
exfever_7980
|
Shawn Carlson is an American scientist and science writer who specializes in the field of physics.
|
{
"evidence": [
"Shawn Carlson",
"Physicist"
],
"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:\nShawn Carlson is an American scientist and science writer who specializes in the field of physics.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Shawn Carlson",
"Physicist"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 3523,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
56824
|
Chumatlán which produces maize and coffee is cold all year and it is delimited to the north by Coyutla.
|
{
"evidence": [
"Chumatlán_sentence_6",
"Chumatlán_sentence_4",
"Chumatlán_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:\nChumatlán which produces maize and coffee is cold all year and it is delimited to the north by Coyutla.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Chumatlán_sentence_6",
"Chumatlán_sentence_4",
"Chumatlán_sentence_5"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 3524,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.621",
"value": "?search="
},
{
"operation": "search",
"time": "48.249",
"value": "Chumatlán"
},
{
"operation": "Now on",
"time": "49.061",
"value": "Chumatlán"
},
{
"operation": "Highlighting",
"time": "71.745",
"value": "Chumatlán_sentence_6"
},
{
"operation": "Highlighting",
"time": "75.312",
"value": "Chumatlán_sentence_4"
},
{
"operation": "Highlighting",
"time": "78.615",
"value": "Chumatlán_sentence_5"
},
{
"operation": "finish",
"time": "93.079",
"value": "finish"
}
]
}
|
exfever_8442
|
Rebecca Maria Hall is an actress who made her first on-screen appearance in The Camomile Lawn, directed by her English theatre, opera, and film director father, who founded the major British theatre company in Stratford-upon-Avon, Warwickshire, England.
|
{
"evidence": [
"Rebecca Hall",
"Peter Hall (director)",
"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:\nRebecca Maria Hall is an actress who made her first on-screen appearance in The Camomile Lawn, directed by her English theatre, opera, and film director father, who founded the major British theatre company in Stratford-upon-Avon, Warwickshire, England.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Rebecca Hall",
"Peter Hall (director)",
"Royal Shakespeare Company"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 3525,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_7171
|
Splash is a romantic movie that became first in the chart of romance films in 1980s.
|
{
"evidence": [
"Splash (film)",
"Daryl Hannah",
"Ridley Scott"
],
"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:\nSplash is a romantic movie that became first in the chart of romance films in 1980s.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Splash (film)",
"Daryl Hannah",
"Ridley Scott"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 3526,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_2118
|
Nickelodeon began life as a test broadcast on December 1, 1987, as part of QUBE, an early cable television system broadcast locally in the 14th-most populous city in the U.S.
|
{
"evidence": [
"Nickelodeon",
"Columbus, Ohio"
],
"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:\nNickelodeon began life as a test broadcast on December 1, 1987, as part of QUBE, an early cable television system broadcast locally in the 14th-most populous city in the U.S.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Nickelodeon",
"Columbus, Ohio"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 3527,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_3165
|
"Inside Man" was the last film directed by Spike Lee.
|
{
"evidence": [
"Inside Man",
"Power broker (politics)"
],
"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:\n\"Inside Man\" was the last film directed by Spike Lee.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Inside Man",
"Power broker (politics)"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 3528,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_4173
|
Colin Quin is an American stand-up comedian and actor known for comedic films like the 2010 American comedy film directed by an American director known for his partnership with comedic actor Adam Sandler.
|
{
"evidence": [
"Colin Quinn",
"Grown Ups (film)",
"Dennis Dugan"
],
"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:\nColin Quin is an American stand-up comedian and actor known for comedic films like the 2010 American comedy film directed by an American director known for his partnership with comedic actor Adam Sandler.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Colin Quinn",
"Grown Ups (film)",
"Dennis Dugan"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 3529,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_3456
|
On October 24, 2015, Sundar Pichai became the second CEO of Google, replacing Larry Page.
|
{
"evidence": [
"Google",
"Cloud computing"
],
"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:\nOn October 24, 2015, Sundar Pichai became the second CEO of Google, replacing Larry Page.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Google",
"Cloud computing"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 3530,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
63138
|
The scientific classification of Uxuma is kingdom animalia, phylum arthropoda and class arachnida but have no legs.
|
{
"evidence": [
"Uxuma_sentence_0",
"Jumping spider_sentence_11"
],
"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 scientific classification of Uxuma is kingdom animalia, phylum arthropoda and class arachnida but have no legs.\n",
"role": "user"
}
] |
Multi-hop Reasoning
|
{
"ground_truth": {
"evidence": [
"Uxuma_sentence_0",
"Jumping spider_sentence_11"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 3531,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "2.493",
"value": "?search="
},
{
"operation": "search",
"time": "14.585",
"value": "Uxuma"
},
{
"operation": "Now on",
"time": "15.289",
"value": "Uxuma"
},
{
"operation": "Highlighting",
"time": "34.123",
"value": "Uxuma_cell_0_2_0"
},
{
"operation": "Highlighting",
"time": "34.127",
"value": "Uxuma_header_cell_0_1_0"
},
{
"operation": "Highlighting",
"time": "34.128",
"value": "Uxuma_header_cell_0_0_0"
},
{
"operation": "Highlighting",
"time": "35.224",
"value": "Uxuma_cell_0_2_1"
},
{
"operation": "Highlighting",
"time": "37.455",
"value": "Uxuma_cell_0_3_0"
},
{
"operation": "Highlighting",
"time": "38.34",
"value": "Uxuma_cell_0_3_1"
},
{
"operation": "Highlighting",
"time": "41.927",
"value": "Uxuma_cell_0_5_0"
},
{
"operation": "Highlighting",
"time": "43.173",
"value": "Uxuma_cell_0_5_1"
},
{
"operation": "Highlighting deleted",
"time": "52.757",
"value": "Uxuma_cell_0_5_1"
},
{
"operation": "Highlighting deleted",
"time": "53.573",
"value": "Uxuma_cell_0_5_0"
},
{
"operation": "Highlighting deleted",
"time": "54.231",
"value": "Uxuma_cell_0_3_1"
},
{
"operation": "Highlighting deleted",
"time": "54.925",
"value": "Uxuma_cell_0_3_0"
},
{
"operation": "Highlighting deleted",
"time": "55.568",
"value": "Uxuma_cell_0_2_1"
},
{
"operation": "Highlighting deleted",
"time": "56.144",
"value": "Uxuma_cell_0_2_0"
},
{
"operation": "Highlighting deleted",
"time": "56.147",
"value": "Uxuma_header_cell_0_1_0"
},
{
"operation": "Highlighting deleted",
"time": "56.148",
"value": "Uxuma_header_cell_0_0_0"
},
{
"operation": "Highlighting",
"time": "63.305",
"value": "Uxuma_sentence_0"
},
{
"operation": "hyperlink",
"time": "65.068",
"value": "jumping spiders"
},
{
"operation": "Now on",
"time": "66.587",
"value": "Jumping spider"
},
{
"operation": "Highlighting",
"time": "120.753",
"value": "Jumping spider_sentence_10"
},
{
"operation": "Highlighting deleted",
"time": "126.44",
"value": "Jumping spider_sentence_10"
},
{
"operation": "Highlighting",
"time": "127.52",
"value": "Jumping spider_sentence_11"
},
{
"operation": "finish",
"time": "148.134",
"value": "finish"
}
]
}
|
exfever_4966
|
Pineapple Express is a 2008 African stoner comedy film written by Evan Goldberg and a Canadian actor who was originally a stand-up comedian in Vancouver.
|
{
"evidence": [
"Pineapple Express (film)",
"Seth Rogen"
],
"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:\nPineapple Express is a 2008 African stoner comedy film written by Evan Goldberg and a Canadian actor who was originally a stand-up comedian in Vancouver.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Pineapple Express (film)",
"Seth Rogen"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 3532,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
30007
|
The first phase of the plan to increase seating capacity for Williams' Stadium to 30,000 expanded the stadium by 7,200 seats with the addition of a second deck to the home side of the stadium.
|
{
"evidence": [
"Williams Stadium_sentence_11",
"Williams Stadium_sentence_12",
"Williams Stadium_sentence_13"
],
"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 first phase of the plan to increase seating capacity for Williams' Stadium to 30,000 expanded the stadium by 7,200 seats with the addition of a second deck to the home side of the stadium.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Williams Stadium_sentence_11",
"Williams Stadium_sentence_12",
"Williams Stadium_sentence_13"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 3533,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.988",
"value": "?search="
},
{
"operation": "search",
"time": "28.947",
"value": "Williams Stadium"
},
{
"operation": "Now on",
"time": "29.898",
"value": "Williams Stadium"
},
{
"operation": "Highlighting",
"time": "56.474",
"value": "Williams Stadium_sentence_11"
},
{
"operation": "Highlighting",
"time": "57.709",
"value": "Williams Stadium_sentence_12"
},
{
"operation": "Highlighting",
"time": "70.579",
"value": "Williams Stadium_sentence_13"
},
{
"operation": "finish",
"time": "95.341",
"value": "finish"
}
]
}
|
exfever_7421
|
Year Zero is the fourth studio album by American industrial rock band Nine Inch Nails, released by an American record label owned by Universal Music Group.
|
{
"evidence": [
"Year Zero (album)",
"Interscope 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:\nYear Zero is the fourth studio album by American industrial rock band Nine Inch Nails, released by an American record label owned by Universal Music Group.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Year Zero (album)",
"Interscope Records"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 3534,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
19362
|
Joséphine Fodor married the actor Jean-Baptiste Tharaud-Mainvielle in 1812, left St. Petersburg during the Great Patriotic War, and moved to France.
|
{
"evidence": [
"Joséphine Fodor_sentence_6",
"French invasion of Russia_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:\nJoséphine Fodor married the actor Jean-Baptiste Tharaud-Mainvielle in 1812, left St. Petersburg during the Great Patriotic War, and moved to France.\n",
"role": "user"
}
] |
Multi-hop Reasoning
|
{
"ground_truth": {
"evidence": [
"Joséphine Fodor_sentence_6",
"French invasion of Russia_sentence_3"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 3535,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.133",
"value": "?search="
},
{
"operation": "search",
"time": "17.059",
"value": "Joséphine Fodor"
},
{
"operation": "Now on",
"time": "18.361",
"value": "Joséphine Fodor"
},
{
"operation": "Highlighting",
"time": "46.109",
"value": "Joséphine Fodor_sentence_6"
},
{
"operation": "Highlighting",
"time": "51.02",
"value": "Joséphine Fodor_sentence_5"
},
{
"operation": "Page search",
"time": "75.738",
"value": "Patriotic"
},
{
"operation": "hyperlink",
"time": "87.739",
"value": "French invasion of Russia"
},
{
"operation": "Now on",
"time": "89.369",
"value": "French invasion of Russia"
},
{
"operation": "Highlighting",
"time": "94.102",
"value": "French invasion of Russia_sentence_3"
},
{
"operation": "finish",
"time": "142.529",
"value": "finish"
}
]
}
|
87976
|
During the Waddington By-Election 17 October 2002, Conservatives won more votes than Liberal Democrats which was founded in 1988.
|
{
"evidence": [
"Lincoln, England_sentence_2"
],
"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:\nDuring the Waddington By-Election 17 October 2002, Conservatives won more votes than Liberal Democrats which was founded in 1988.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Lincoln, England_sentence_2"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 3536,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.387",
"value": "undefined"
},
{
"operation": "search",
"time": "27.017",
"value": "Waddington, Lancashire"
},
{
"operation": "Now on",
"time": "28.609",
"value": "Waddington, Lancashire"
},
{
"operation": "search",
"time": "99.418",
"value": "Waddington By-Election"
},
{
"operation": "Now on",
"time": "100.544",
"value": "undefined"
},
{
"operation": "hyperlink",
"time": "121.211",
"value": "Lincolnshire"
},
{
"operation": "Now on",
"time": "124.816",
"value": "Lincolnshire"
},
{
"operation": "Page search",
"time": "130.541",
"value": "waddin"
},
{
"operation": "Page search",
"time": "130.813",
"value": "waddin"
},
{
"operation": "Page search",
"time": "131.061",
"value": "waddin"
},
{
"operation": "Page search",
"time": "131.314",
"value": "waddin"
},
{
"operation": "Page search",
"time": "131.607",
"value": "waddin"
},
{
"operation": "Page search",
"time": "131.858",
"value": "waddin"
},
{
"operation": "Page search",
"time": "132.098",
"value": "waddin"
},
{
"operation": "back-button-clicked",
"time": "163.073",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "164.384",
"value": "undefined"
},
{
"operation": "search",
"time": "199.017",
"value": "2002 Waddington Election"
},
{
"operation": "Now on",
"time": "200.24",
"value": "undefined"
},
{
"operation": "hyperlink",
"time": "215.899",
"value": "Lincoln, England"
},
{
"operation": "Now on",
"time": "217.91",
"value": "Lincoln, England"
},
{
"operation": "Page search",
"time": "224.687",
"value": "waddin"
},
{
"operation": "Page search",
"time": "224.958",
"value": "waddin"
},
{
"operation": "Highlighting",
"time": "228.077",
"value": "Lincoln, England_sentence_2"
},
{
"operation": "hyperlink",
"time": "229.808",
"value": "Waddington"
},
{
"operation": "Now on",
"time": "230.872",
"value": "Waddington, Lincolnshire"
},
{
"operation": "finish",
"time": "249.763",
"value": "finish"
}
]
}
|
exfever_2802
|
Sarah Louise Palin was Alaska's shortest-serving governor.
|
{
"evidence": [
"Sarah Palin",
"Major party"
],
"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:\nSarah Louise Palin was Alaska's shortest-serving governor.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Sarah Palin",
"Major party"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 3537,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
36418
|
John Beck had more wins than Daunte Culpepper.
|
{
"evidence": [
"Daunte Culpepper_sentence_0",
"John Beck (gridiron football)_sentence_1",
"John Beck (gridiron football)_sentence_2"
],
"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:\nJohn Beck had more wins than Daunte Culpepper.\n",
"role": "user"
}
] |
Multi-hop Reasoning
|
{
"ground_truth": {
"evidence": [
"Daunte Culpepper_sentence_0",
"John Beck (gridiron football)_sentence_1",
"John Beck (gridiron football)_sentence_2"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 3538,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.383",
"value": "?search="
},
{
"operation": "search",
"time": "41.303",
"value": "John Beck"
},
{
"operation": "Now on",
"time": "42.261",
"value": "John Beck"
},
{
"operation": "hyperlink",
"time": "50.674",
"value": "John Beck (gridiron football)"
},
{
"operation": "Now on",
"time": "51.476",
"value": "John Beck (gridiron football)"
},
{
"operation": "Page search",
"time": "59.627",
"value": "Daunte Culpepper"
},
{
"operation": "back-button-clicked",
"time": "71.537",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "71.609",
"value": "John Beck"
},
{
"operation": "search",
"time": "84.434",
"value": "Daunte Culpepper"
},
{
"operation": "Now on",
"time": "86.183",
"value": "Daunte Culpepper"
},
{
"operation": "Highlighting",
"time": "89.162",
"value": "Daunte Culpepper_sentence_0"
},
{
"operation": "Page search",
"time": "110.497",
"value": "John Beck"
},
{
"operation": "search",
"time": "133.376",
"value": "John Beck had more wins than Daunte Culpepper."
},
{
"operation": "Now on",
"time": "134.898",
"value": "undefined"
},
{
"operation": "hyperlink",
"time": "139.379",
"value": "2007 Miami Dolphins season"
},
{
"operation": "Now on",
"time": "140.787",
"value": "2007 Miami Dolphins season"
},
{
"operation": "Page search",
"time": "152.218",
"value": "john beck"
},
{
"operation": "Page search",
"time": "156.888",
"value": "john beck"
},
{
"operation": "hyperlink",
"time": "174.258",
"value": "John Beck"
},
{
"operation": "Now on",
"time": "175.037",
"value": "John Beck (American football player)"
},
{
"operation": "Highlighting",
"time": "176.654",
"value": "John Beck (gridiron football)_sentence_1"
},
{
"operation": "Highlighting",
"time": "177.848",
"value": "John Beck (gridiron football)_sentence_2"
},
{
"operation": "back-button-clicked",
"time": "241.199",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "241.437",
"value": "2007 Miami Dolphins season"
},
{
"operation": "finish",
"time": "294.961",
"value": "finish"
}
]
}
|
57582
|
Captaincy of the Portuguese Empire were developed successively, Prince Henry the Navigator instituted the Captaincy system and remained on the mainland.
|
{
"evidence": [
"Captaincy_sentence_5",
"Captaincy_sentence_6",
"Captaincy_sentence_7"
],
"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:\nCaptaincy of the Portuguese Empire were developed successively, Prince Henry the Navigator instituted the Captaincy system and remained on the mainland.\n",
"role": "user"
}
] |
Entity Disambiguation
|
{
"ground_truth": {
"evidence": [
"Captaincy_sentence_5",
"Captaincy_sentence_6",
"Captaincy_sentence_7"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 3539,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.831",
"value": "?search="
},
{
"operation": "search",
"time": "11.51",
"value": "Contains...Captaincy of the Portuguese Empire"
},
{
"operation": "Now on",
"time": "13.269",
"value": "undefined"
},
{
"operation": "hyperlink",
"time": "19.171",
"value": "Captaincies of the Portuguese Empire"
},
{
"operation": "Now on",
"time": "19.69",
"value": "Captaincies of the Portuguese Empire"
},
{
"operation": "Highlighting",
"time": "45.018",
"value": "Captaincies of the Portuguese Empire_sentence_0"
},
{
"operation": "Highlighting",
"time": "45.019",
"value": "Captaincies of the Portuguese Empire_title"
},
{
"operation": "hyperlink",
"time": "50.138",
"value": "Henry the Navigator"
},
{
"operation": "Now on",
"time": "51.523",
"value": "Prince Henry the Navigator"
},
{
"operation": "search",
"time": "109.006",
"value": "Captaincy"
},
{
"operation": "Now on",
"time": "109.681",
"value": "Captaincy"
},
{
"operation": "Highlighting",
"time": "120.537",
"value": "Captaincy_sentence_5"
},
{
"operation": "Highlighting",
"time": "127.154",
"value": "Captaincy_sentence_6"
},
{
"operation": "Highlighting",
"time": "151.001",
"value": "Captaincy_sentence_7"
},
{
"operation": "finish",
"time": "182.531",
"value": "finish"
}
]
}
|
31130
|
The song begins with a sample of the "put your bodies upon the gears" speech address given by Mario Savio, a key figure in the Berkeley Free Speech Movement, which bear some traces of a commentary on the abuses of the powerful.
|
{
"evidence": [
"Wretches and Kings_sentence_6",
"Wretches and Kings_sentence_7"
],
"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 song begins with a sample of the \"put your bodies upon the gears\" speech address given by Mario Savio, a key figure in the Berkeley Free Speech Movement, which bear some traces of a commentary on the abuses of the powerful.\n",
"role": "user"
}
] |
Search terms not in claim
|
{
"ground_truth": {
"evidence": [
"Wretches and Kings_sentence_6",
"Wretches and Kings_sentence_7"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 3540,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.486",
"value": "undefined"
},
{
"operation": "search",
"time": "97.416",
"value": "Mario Savio"
},
{
"operation": "Now on",
"time": "98.159",
"value": "Mario Savio"
},
{
"operation": "Page search",
"time": "111.56",
"value": "\"put your bodies upon the gears\""
},
{
"operation": "search",
"time": "142.942",
"value": "\"put your bodies upon the gears\""
},
{
"operation": "Now on",
"time": "143.553",
"value": "undefined"
},
{
"operation": "hyperlink",
"time": "148.458",
"value": "Wretches and Kings"
},
{
"operation": "Now on",
"time": "148.877",
"value": "Wretches and Kings"
},
{
"operation": "Page search",
"time": "153.213",
"value": "\"put your bodies upon the gears\""
},
{
"operation": "Highlighting",
"time": "155.88",
"value": "Wretches and Kings_sentence_6"
},
{
"operation": "Highlighting",
"time": "165.23",
"value": "Wretches and Kings_sentence_7"
},
{
"operation": "finish",
"time": "294.117",
"value": "finish"
}
]
}
|
94375
|
The outside of Barbour County Courthouse is made of Hummelstown brownstone, a dark brown with purplish hues.
|
{
"evidence": [
"Barbour County Courthouse_sentence_13",
"Barbour County Courthouse_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:\nThe outside of Barbour County Courthouse is made of Hummelstown brownstone, a dark brown with purplish hues.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Barbour County Courthouse_sentence_13",
"Barbour County Courthouse_sentence_15"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 3541,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.056",
"value": "?search="
},
{
"operation": "search",
"time": "53.453",
"value": "Barbour County Courthouse"
},
{
"operation": "Now on",
"time": "54.123",
"value": "Barbour County Courthouse"
},
{
"operation": "hyperlink",
"time": "76.011",
"value": "2 Description Barbour County Courthouse_section_3"
},
{
"operation": "Highlighting",
"time": "80.915",
"value": "Barbour County Courthouse_sentence_13"
},
{
"operation": "Highlighting",
"time": "103.595",
"value": "Barbour County Courthouse_sentence_15"
},
{
"operation": "finish",
"time": "153.044",
"value": "finish"
}
]
}
|
exfever_7279
|
Daniel Bergman is a cyberpunk film directed by Robert Longo starring a Swedish actor who starred in Rocky IV as the imposing Soviet boxer Ivan Drago.
|
{
"evidence": [
"Johnny Mnemonic (film)",
"Dolph Lundgren"
],
"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:\nDaniel Bergman is a cyberpunk film directed by Robert Longo starring a Swedish actor who starred in Rocky IV as the imposing Soviet boxer Ivan Drago.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Johnny Mnemonic (film)",
"Dolph Lundgren"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 3542,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
40176
|
In the girls badminton at the 2018 Summer Youth Olympics, Wang Zhiyi of China was ranked higher than Wei Yaxin, also of China.
|
{
"evidence": [
"Wang Zhiyi_sentence_1",
"Wang Zhiyi_sentence_6",
"Wei Yaxin_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:\nIn the girls badminton at the 2018 Summer Youth Olympics, Wang Zhiyi of China was ranked higher than Wei Yaxin, also of China.\n",
"role": "user"
}
] |
Multi-hop Reasoning
|
{
"ground_truth": {
"evidence": [
"Wang Zhiyi_sentence_1",
"Wang Zhiyi_sentence_6",
"Wei Yaxin_sentence_1"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 3543,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.003",
"value": "?search="
},
{
"operation": "search",
"time": "15.961",
"value": "2018 Summer Youth Olympics"
},
{
"operation": "Now on",
"time": "17.151",
"value": "2018 Summer Youth Olympics"
},
{
"operation": "search",
"time": "32.638",
"value": "Contains...girls badminton at the 2018 summer olymics"
},
{
"operation": "Now on",
"time": "33.556",
"value": "undefined"
},
{
"operation": "search",
"time": "65.431",
"value": "Badminton at the 2018 Summer Youth Olympics – Girls' singles"
},
{
"operation": "Now on",
"time": "66.81",
"value": "Badminton at the 2018 Summer Youth Olympics – Girls' singles"
},
{
"operation": "Highlighting",
"time": "107.81",
"value": "Badminton at the 2018 Summer Youth Olympics – Girls' singles_header_cell_8_0_0"
},
{
"operation": "Highlighting",
"time": "107.811",
"value": "Badminton at the 2018 Summer Youth Olympics – Girls' singles_title"
},
{
"operation": "Highlighting",
"time": "107.812",
"value": "Badminton at the 2018 Summer Youth Olympics – Girls' singles_section_1"
},
{
"operation": "Highlighting",
"time": "107.813",
"value": "Badminton at the 2018 Summer Youth Olympics – Girls' singles_section_2"
},
{
"operation": "Highlighting",
"time": "107.813",
"value": "Badminton at the 2018 Summer Youth Olympics – Girls' singles_section_10"
},
{
"operation": "Highlighting",
"time": "109.994",
"value": "Badminton at the 2018 Summer Youth Olympics – Girls' singles_cell_8_2_0"
},
{
"operation": "Highlighting",
"time": "110",
"value": "Badminton at the 2018 Summer Youth Olympics – Girls' singles_header_cell_8_0_0"
},
{
"operation": "Highlighting",
"time": "111.858",
"value": "Badminton at the 2018 Summer Youth Olympics – Girls' singles_header_cell_8_0_1"
},
{
"operation": "Highlighting",
"time": "113.631",
"value": "Badminton at the 2018 Summer Youth Olympics – Girls' singles_header_cell_8_1_0"
},
{
"operation": "Highlighting",
"time": "114.231",
"value": "Badminton at the 2018 Summer Youth Olympics – Girls' singles_cell_8_2_1"
},
{
"operation": "Highlighting",
"time": "114.236",
"value": "Badminton at the 2018 Summer Youth Olympics – Girls' singles_header_cell_8_1_0"
},
{
"operation": "Highlighting",
"time": "114.24",
"value": "Badminton at the 2018 Summer Youth Olympics – Girls' singles_header_cell_8_0_1"
},
{
"operation": "Highlighting",
"time": "114.818",
"value": "Badminton at the 2018 Summer Youth Olympics – Girls' singles_header_cell_8_1_1"
},
{
"operation": "Highlighting",
"time": "115.329",
"value": "Badminton at the 2018 Summer Youth Olympics – Girls' singles_cell_8_2_2"
},
{
"operation": "Highlighting",
"time": "115.337",
"value": "Badminton at the 2018 Summer Youth Olympics – Girls' singles_header_cell_8_1_1"
},
{
"operation": "Highlighting",
"time": "115.866",
"value": "Badminton at the 2018 Summer Youth Olympics – Girls' singles_header_cell_8_1_2"
},
{
"operation": "Highlighting",
"time": "116.488",
"value": "Badminton at the 2018 Summer Youth Olympics – Girls' singles_cell_8_2_3"
},
{
"operation": "Highlighting",
"time": "116.497",
"value": "Badminton at the 2018 Summer Youth Olympics – Girls' singles_header_cell_8_1_2"
},
{
"operation": "Highlighting deleted",
"time": "126.059",
"value": "Badminton at the 2018 Summer Youth Olympics – Girls' singles_cell_8_2_1"
},
{
"operation": "Highlighting deleted",
"time": "126.061",
"value": "Badminton at the 2018 Summer Youth Olympics – Girls' singles_header_cell_8_1_0"
},
{
"operation": "Highlighting deleted",
"time": "126.673",
"value": "Badminton at the 2018 Summer Youth Olympics – Girls' singles_cell_8_2_2"
},
{
"operation": "Highlighting deleted",
"time": "126.68",
"value": "Badminton at the 2018 Summer Youth Olympics – Girls' singles_header_cell_8_1_1"
},
{
"operation": "Highlighting deleted",
"time": "127.141",
"value": "Badminton at the 2018 Summer Youth Olympics – Girls' singles_cell_8_2_3"
},
{
"operation": "Highlighting deleted",
"time": "127.146",
"value": "Badminton at the 2018 Summer Youth Olympics – Girls' singles_header_cell_8_1_2"
},
{
"operation": "Highlighting deleted",
"time": "127.147",
"value": "Badminton at the 2018 Summer Youth Olympics – Girls' singles_header_cell_8_0_1"
},
{
"operation": "Highlighting deleted",
"time": "128.167",
"value": "Badminton at the 2018 Summer Youth Olympics – Girls' singles_cell_8_2_0"
},
{
"operation": "Highlighting deleted",
"time": "128.169",
"value": "Badminton at the 2018 Summer Youth Olympics – Girls' singles_title"
},
{
"operation": "Highlighting deleted",
"time": "128.172",
"value": "Badminton at the 2018 Summer Youth Olympics – Girls' singles_section_1"
},
{
"operation": "Highlighting deleted",
"time": "128.174",
"value": "Badminton at the 2018 Summer Youth Olympics – Girls' singles_section_2"
},
{
"operation": "Highlighting deleted",
"time": "128.175",
"value": "Badminton at the 2018 Summer Youth Olympics – Girls' singles_section_10"
},
{
"operation": "Highlighting deleted",
"time": "128.177",
"value": "Badminton at the 2018 Summer Youth Olympics – Girls' singles_header_cell_8_0_0"
},
{
"operation": "Highlighting",
"time": "170.762",
"value": "Badminton at the 2018 Summer Youth Olympics – Girls' singles_header_cell_8_0_0"
},
{
"operation": "Highlighting",
"time": "170.764",
"value": "Badminton at the 2018 Summer Youth Olympics – Girls' singles_title"
},
{
"operation": "Highlighting",
"time": "170.765",
"value": "Badminton at the 2018 Summer Youth Olympics – Girls' singles_section_1"
},
{
"operation": "Highlighting",
"time": "170.765",
"value": "Badminton at the 2018 Summer Youth Olympics – Girls' singles_section_2"
},
{
"operation": "Highlighting",
"time": "170.766",
"value": "Badminton at the 2018 Summer Youth Olympics – Girls' singles_section_10"
},
{
"operation": "Highlighting",
"time": "172.061",
"value": "Badminton at the 2018 Summer Youth Olympics – Girls' singles_cell_8_2_0"
},
{
"operation": "Highlighting",
"time": "172.067",
"value": "Badminton at the 2018 Summer Youth Olympics – Girls' singles_header_cell_8_0_0"
},
{
"operation": "Highlighting",
"time": "173.297",
"value": "Badminton at the 2018 Summer Youth Olympics – Girls' singles_header_cell_8_0_1"
},
{
"operation": "Highlighting",
"time": "174.038",
"value": "Badminton at the 2018 Summer Youth Olympics – Girls' singles_header_cell_8_1_0"
},
{
"operation": "Highlighting deleted",
"time": "175.137",
"value": "Badminton at the 2018 Summer Youth Olympics – Girls' singles_header_cell_8_1_0"
},
{
"operation": "Highlighting deleted",
"time": "175.713",
"value": "Badminton at the 2018 Summer Youth Olympics – Girls' singles_header_cell_8_0_1"
},
{
"operation": "Highlighting",
"time": "176.322",
"value": "Badminton at the 2018 Summer Youth Olympics – Girls' singles_header_cell_8_0_1"
},
{
"operation": "Highlighting",
"time": "176.809",
"value": "Badminton at the 2018 Summer Youth Olympics – Girls' singles_header_cell_8_1_0"
},
{
"operation": "Highlighting",
"time": "177.285",
"value": "Badminton at the 2018 Summer Youth Olympics – Girls' singles_cell_8_2_1"
},
{
"operation": "Highlighting",
"time": "177.292",
"value": "Badminton at the 2018 Summer Youth Olympics – Girls' singles_header_cell_8_1_0"
},
{
"operation": "Highlighting",
"time": "177.295",
"value": "Badminton at the 2018 Summer Youth Olympics – Girls' singles_header_cell_8_0_1"
},
{
"operation": "Highlighting",
"time": "177.764",
"value": "Badminton at the 2018 Summer Youth Olympics – Girls' singles_header_cell_8_1_1"
},
{
"operation": "Highlighting",
"time": "178.22",
"value": "Badminton at the 2018 Summer Youth Olympics – Girls' singles_cell_8_2_2"
},
{
"operation": "Highlighting",
"time": "178.23",
"value": "Badminton at the 2018 Summer Youth Olympics – Girls' singles_header_cell_8_1_1"
},
{
"operation": "Highlighting",
"time": "178.765",
"value": "Badminton at the 2018 Summer Youth Olympics – Girls' singles_header_cell_8_1_2"
},
{
"operation": "Highlighting",
"time": "179.254",
"value": "Badminton at the 2018 Summer Youth Olympics – Girls' singles_cell_8_3_3"
},
{
"operation": "Highlighting",
"time": "179.263",
"value": "Badminton at the 2018 Summer Youth Olympics – Girls' singles_header_cell_8_1_2"
},
{
"operation": "hyperlink",
"time": "180.167",
"value": "CHN"
},
{
"operation": "Now on",
"time": "181.719",
"value": "China at the 2018 Summer Youth Olympics"
},
{
"operation": "back-button-clicked",
"time": "182.064",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "182.196",
"value": "Badminton at the 2018 Summer Youth Olympics – Girls' singles"
},
{
"operation": "hyperlink",
"time": "183.595",
"value": "Wang Zhiyi"
},
{
"operation": "Now on",
"time": "184.154",
"value": "Wang Zhiyi"
},
{
"operation": "Highlighting",
"time": "202.935",
"value": "Wang Zhiyi_sentence_1"
},
{
"operation": "Highlighting",
"time": "206.617",
"value": "Wang Zhiyi_sentence_6"
},
{
"operation": "Highlighting",
"time": "227.347",
"value": "Wang Zhiyi_sentence_1"
},
{
"operation": "Highlighting",
"time": "228.815",
"value": "Wang Zhiyi_sentence_6"
},
{
"operation": "search",
"time": "237.634",
"value": "Wei Yaxin"
},
{
"operation": "Now on",
"time": "238.359",
"value": "Wei Yaxin"
},
{
"operation": "Highlighting",
"time": "250.283",
"value": "Wei Yaxin_sentence_1"
},
{
"operation": "back-button-clicked",
"time": "277.807",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "277.881",
"value": "Wang Zhiyi"
},
{
"operation": "back-button-clicked",
"time": "279.601",
"value": "back-button-clicked"
},
{
"operation": "back-button-clicked",
"time": "279.606",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "279.705",
"value": "Badminton at the 2018 Summer Youth Olympics – Girls' singles"
},
{
"operation": "Highlighting",
"time": "282.558",
"value": "Badminton at the 2018 Summer Youth Olympics – Girls' singles_header_cell_8_0_0"
},
{
"operation": "Highlighting",
"time": "282.56",
"value": "Badminton at the 2018 Summer Youth Olympics – Girls' singles_title"
},
{
"operation": "Highlighting",
"time": "282.561",
"value": "Badminton at the 2018 Summer Youth Olympics – Girls' singles_section_1"
},
{
"operation": "Highlighting",
"time": "282.561",
"value": "Badminton at the 2018 Summer Youth Olympics – Girls' singles_section_2"
},
{
"operation": "Highlighting",
"time": "282.562",
"value": "Badminton at the 2018 Summer Youth Olympics – Girls' singles_section_10"
},
{
"operation": "Highlighting deleted",
"time": "283.352",
"value": "Badminton at the 2018 Summer Youth Olympics – Girls' singles_header_cell_8_0_0"
},
{
"operation": "Highlighting deleted",
"time": "283.352",
"value": "Badminton at the 2018 Summer Youth Olympics – Girls' singles_title"
},
{
"operation": "Highlighting deleted",
"time": "283.354",
"value": "Badminton at the 2018 Summer Youth Olympics – Girls' singles_section_1"
},
{
"operation": "Highlighting deleted",
"time": "283.355",
"value": "Badminton at the 2018 Summer Youth Olympics – Girls' singles_section_2"
},
{
"operation": "Highlighting deleted",
"time": "283.356",
"value": "Badminton at the 2018 Summer Youth Olympics – Girls' singles_section_10"
},
{
"operation": "Highlighting",
"time": "284.266",
"value": "Badminton at the 2018 Summer Youth Olympics – Girls' singles_header_cell_8_0_0"
},
{
"operation": "Highlighting",
"time": "284.268",
"value": "Badminton at the 2018 Summer Youth Olympics – Girls' singles_title"
},
{
"operation": "Highlighting",
"time": "284.268",
"value": "Badminton at the 2018 Summer Youth Olympics – Girls' singles_section_1"
},
{
"operation": "Highlighting",
"time": "284.268",
"value": "Badminton at the 2018 Summer Youth Olympics – Girls' singles_section_2"
},
{
"operation": "Highlighting",
"time": "284.269",
"value": "Badminton at the 2018 Summer Youth Olympics – Girls' singles_section_10"
},
{
"operation": "Highlighting",
"time": "284.832",
"value": "Badminton at the 2018 Summer Youth Olympics – Girls' singles_cell_8_2_0"
},
{
"operation": "Highlighting",
"time": "284.841",
"value": "Badminton at the 2018 Summer Youth Olympics – Girls' singles_header_cell_8_0_0"
},
{
"operation": "finish",
"time": "380.592",
"value": "finish"
}
]
}
|
exfever_3603
|
Hotel Transylvania 2 was animated by a Canadian visual effects and computer animation studio that received Oscars for its work on a 2004 American superhero film directed by Sam Raimi.
|
{
"evidence": [
"Hotel Transylvania 2",
"Sony Pictures Imageworks",
"Spider-Man 2"
],
"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:\nHotel Transylvania 2 was animated by a Canadian visual effects and computer animation studio that received Oscars for its work on a 2004 American superhero film directed by Sam Raimi.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Hotel Transylvania 2",
"Sony Pictures Imageworks",
"Spider-Man 2"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 3544,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
48028
|
3133 Sendai was discovered by Albert Smith on September 24, 1907.
|
{
"evidence": [
"3133 Sendai_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:\n3133 Sendai was discovered by Albert Smith on September 24, 1907.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"3133 Sendai_sentence_1"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 3545,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.91",
"value": "?search="
},
{
"operation": "search",
"time": "13.862",
"value": "3133 Sendai"
},
{
"operation": "Now on",
"time": "14.592",
"value": "3133 Sendai"
},
{
"operation": "Highlighting",
"time": "22.998",
"value": "3133 Sendai_sentence_1"
},
{
"operation": "hyperlink",
"time": "111.412",
"value": "A. Kopff"
},
{
"operation": "Now on",
"time": "112.052",
"value": "August Kopff"
},
{
"operation": "finish",
"time": "139.083",
"value": "finish"
}
]
}
|
18984
|
When designing the synagogue, Förster made some innovations: he had an organ built and the Torah reader's platform was put before the Ark, not in the centre of the synagogue; these innovations were accepted by the Orthodox majority of the city's Jews.
|
{
"evidence": [
"Synagogue of Miskolc_sentence_6",
"Synagogue of Miskolc_sentence_7"
],
"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:\nWhen designing the synagogue, Förster made some innovations: he had an organ built and the Torah reader's platform was put before the Ark, not in the centre of the synagogue; these innovations were accepted by the Orthodox majority of the city's Jews.\n",
"role": "user"
}
] |
Search terms not in claim
|
{
"ground_truth": {
"evidence": [
"Synagogue of Miskolc_sentence_6",
"Synagogue of Miskolc_sentence_7"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 3546,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.786",
"value": "?search="
},
{
"operation": "search",
"time": "23.455",
"value": "Förster"
},
{
"operation": "Now on",
"time": "24.488",
"value": "Förster"
},
{
"operation": "hyperlink",
"time": "38.688",
"value": "Ludwig Förster"
},
{
"operation": "Now on",
"time": "39.107",
"value": "Ludwig Förster"
},
{
"operation": "hyperlink",
"time": "63.557",
"value": "Dohány Street Synagogue"
},
{
"operation": "Now on",
"time": "64.426",
"value": "Dohány Street Synagogue"
},
{
"operation": "Page search",
"time": "75.853",
"value": "organ"
},
{
"operation": "back-button-clicked",
"time": "109.952",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "110",
"value": "Ludwig Förster"
},
{
"operation": "Page search",
"time": "115.475",
"value": "organ"
},
{
"operation": "hyperlink",
"time": "143.227",
"value": "Leopoldstädter Tempel"
},
{
"operation": "Now on",
"time": "143.874",
"value": "Leopoldstädter Tempel"
},
{
"operation": "Page search",
"time": "147.114",
"value": "organ"
},
{
"operation": "back-button-clicked",
"time": "155.689",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "155.738",
"value": "Ludwig Förster"
},
{
"operation": "hyperlink",
"time": "160.934",
"value": "Synagogue of Miskolc"
},
{
"operation": "Now on",
"time": "161.363",
"value": "Synagogue of Miskolc"
},
{
"operation": "Page search",
"time": "163.754",
"value": "organ"
},
{
"operation": "Highlighting",
"time": "165.418",
"value": "Synagogue of Miskolc_sentence_6"
},
{
"operation": "Highlighting",
"time": "179.12",
"value": "Synagogue of Miskolc_sentence_7"
},
{
"operation": "finish",
"time": "193.047",
"value": "finish"
}
]
}
|
exfever_4247
|
Stan Lee and Steve Ditko created Spider-Man, a stock character that possesses a superpower.
|
{
"evidence": [
"Spider-Man",
"Superhero"
],
"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:\nStan Lee and Steve Ditko created Spider-Man, a stock character that possesses a superpower.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Spider-Man",
"Superhero"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 3547,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_8790
|
Justin Drew Bieber is the youngest Canadian singer.
|
{
"evidence": [
"Justin Bieber",
"Extended play"
],
"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:\nJustin Drew Bieber is the youngest Canadian singer.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Justin Bieber",
"Extended play"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 3548,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
24943
|
Redhook Ale Brewery produces several styles of beer that are distributed through a network of independent sales consultants; their flagship brand is Redhook ESB and it also has other standard and seasonal beers.
|
{
"evidence": [
"Redhook Ale Brewery_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:\nRedhook Ale Brewery produces several styles of beer that are distributed through a network of independent sales consultants; their flagship brand is Redhook ESB and it also has other standard and seasonal beers.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Redhook Ale Brewery_sentence_6"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 3549,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.877",
"value": "?search="
},
{
"operation": "search",
"time": "10.592",
"value": "Redhook Ale Brewery"
},
{
"operation": "Now on",
"time": "11.631",
"value": "Redhook Ale Brewery"
},
{
"operation": "Highlighting",
"time": "35.532",
"value": "Redhook Ale Brewery_sentence_2"
},
{
"operation": "Highlighting deleted",
"time": "38.341",
"value": "Redhook Ale Brewery_sentence_2"
},
{
"operation": "Highlighting",
"time": "40.459",
"value": "Redhook Ale Brewery_sentence_2"
},
{
"operation": "Page search",
"time": "48.991",
"value": "Redhook"
},
{
"operation": "page-search-reset",
"time": "59.843",
"value": "page-search-reset"
},
{
"operation": "Highlighting",
"time": "102.117",
"value": "Redhook Ale Brewery_sentence_4"
},
{
"operation": "Highlighting",
"time": "129.774",
"value": "Redhook Ale Brewery_sentence_6"
},
{
"operation": "Highlighting deleted",
"time": "132.558",
"value": "Redhook Ale Brewery_sentence_4"
},
{
"operation": "Highlighting deleted",
"time": "146.198",
"value": "Redhook Ale Brewery_sentence_2"
},
{
"operation": "finish",
"time": "159.871",
"value": "finish"
}
]
}
|
35929
|
The list of Air Dolomiti destinations include two Austrian cities, four German cities, and 16 Italian cities including Milan, the most populous city in Italy.
|
{
"evidence": [
"Milan_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:\nThe list of Air Dolomiti destinations include two Austrian cities, four German cities, and 16 Italian cities including Milan, the most populous city in Italy.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Milan_sentence_2"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 3550,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.092",
"value": "?search="
},
{
"operation": "search",
"time": "17.196",
"value": "List of Air Dolomiti destinations"
},
{
"operation": "Now on",
"time": "18.131",
"value": "List of Air Dolomiti destinations"
},
{
"operation": "Highlighting",
"time": "40.817",
"value": "List of Air Dolomiti destinations_cell_0_1_0"
},
{
"operation": "Highlighting",
"time": "40.818",
"value": "List of Air Dolomiti destinations_title"
},
{
"operation": "Highlighting",
"time": "40.818",
"value": "List of Air Dolomiti destinations_section_0"
},
{
"operation": "Highlighting",
"time": "40.819",
"value": "List of Air Dolomiti destinations_header_cell_0_0_0"
},
{
"operation": "Highlighting",
"time": "41.486",
"value": "List of Air Dolomiti destinations_cell_0_1_1"
},
{
"operation": "Highlighting",
"time": "41.489",
"value": "List of Air Dolomiti destinations_header_cell_0_0_1"
},
{
"operation": "Highlighting",
"time": "42.158",
"value": "List of Air Dolomiti destinations_cell_0_2_0"
},
{
"operation": "Highlighting",
"time": "42.878",
"value": "List of Air Dolomiti destinations_cell_0_2_1"
},
{
"operation": "Highlighting",
"time": "45.001",
"value": "List of Air Dolomiti destinations_cell_0_3_0"
},
{
"operation": "Highlighting",
"time": "45.679",
"value": "List of Air Dolomiti destinations_cell_0_3_1"
},
{
"operation": "Highlighting",
"time": "47.223",
"value": "List of Air Dolomiti destinations_cell_0_4_0"
},
{
"operation": "Highlighting",
"time": "48.347",
"value": "List of Air Dolomiti destinations_cell_0_4_1"
},
{
"operation": "Highlighting",
"time": "49.743",
"value": "List of Air Dolomiti destinations_cell_0_5_0"
},
{
"operation": "Highlighting",
"time": "50.716",
"value": "List of Air Dolomiti destinations_cell_0_5_1"
},
{
"operation": "Highlighting",
"time": "51.615",
"value": "List of Air Dolomiti destinations_cell_0_6_0"
},
{
"operation": "Highlighting",
"time": "52.39",
"value": "List of Air Dolomiti destinations_cell_0_6_1"
},
{
"operation": "Highlighting",
"time": "58.548",
"value": "List of Air Dolomiti destinations_cell_0_7_0"
},
{
"operation": "Highlighting",
"time": "60.057",
"value": "List of Air Dolomiti destinations_cell_0_7_1"
},
{
"operation": "Highlighting",
"time": "60.903",
"value": "List of Air Dolomiti destinations_cell_0_8_0"
},
{
"operation": "Highlighting",
"time": "61.616",
"value": "List of Air Dolomiti destinations_cell_0_8_1"
},
{
"operation": "Highlighting",
"time": "62.336",
"value": "List of Air Dolomiti destinations_cell_0_9_0"
},
{
"operation": "Highlighting",
"time": "63.521",
"value": "List of Air Dolomiti destinations_cell_0_9_1"
},
{
"operation": "Highlighting",
"time": "64.375",
"value": "List of Air Dolomiti destinations_cell_0_10_0"
},
{
"operation": "Highlighting",
"time": "65.313",
"value": "List of Air Dolomiti destinations_cell_0_10_1"
},
{
"operation": "Highlighting",
"time": "65.991",
"value": "List of Air Dolomiti destinations_cell_0_11_0"
},
{
"operation": "Highlighting",
"time": "66.957",
"value": "List of Air Dolomiti destinations_cell_0_11_1"
},
{
"operation": "Highlighting",
"time": "67.846",
"value": "List of Air Dolomiti destinations_cell_0_12_0"
},
{
"operation": "Highlighting",
"time": "68.852",
"value": "List of Air Dolomiti destinations_cell_0_12_1"
},
{
"operation": "Highlighting",
"time": "70.066",
"value": "List of Air Dolomiti destinations_cell_0_13_0"
},
{
"operation": "Highlighting",
"time": "71.406",
"value": "List of Air Dolomiti destinations_cell_0_13_1"
},
{
"operation": "Highlighting",
"time": "72.135",
"value": "List of Air Dolomiti destinations_cell_0_14_0"
},
{
"operation": "Highlighting",
"time": "73.043",
"value": "List of Air Dolomiti destinations_cell_0_14_1"
},
{
"operation": "Highlighting",
"time": "73.943",
"value": "List of Air Dolomiti destinations_cell_0_15_0"
},
{
"operation": "Highlighting",
"time": "75.109",
"value": "List of Air Dolomiti destinations_cell_0_15_1"
},
{
"operation": "Highlighting",
"time": "75.933",
"value": "List of Air Dolomiti destinations_cell_0_16_0"
},
{
"operation": "Highlighting",
"time": "76.591",
"value": "List of Air Dolomiti destinations_cell_0_16_1"
},
{
"operation": "hyperlink",
"time": "78.12",
"value": "Milan"
},
{
"operation": "Now on",
"time": "79.392",
"value": "Milan"
},
{
"operation": "Highlighting",
"time": "86.115",
"value": "Milan_sentence_2"
},
{
"operation": "back-button-clicked",
"time": "89.243",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "89.388",
"value": "List of Air Dolomiti destinations"
},
{
"operation": "Highlighting deleted",
"time": "90.78",
"value": "List of Air Dolomiti destinations_cell_0_16_0"
},
{
"operation": "Highlighting deleted",
"time": "91.468",
"value": "List of Air Dolomiti destinations_cell_0_16_1"
},
{
"operation": "Highlighting deleted",
"time": "92.035",
"value": "List of Air Dolomiti destinations_cell_0_15_1"
},
{
"operation": "Highlighting deleted",
"time": "92.572",
"value": "List of Air Dolomiti destinations_cell_0_14_1"
},
{
"operation": "Highlighting deleted",
"time": "93.199",
"value": "List of Air Dolomiti destinations_cell_0_13_1"
},
{
"operation": "Highlighting deleted",
"time": "93.802",
"value": "List of Air Dolomiti destinations_cell_0_12_1"
},
{
"operation": "Highlighting deleted",
"time": "94.388",
"value": "List of Air Dolomiti destinations_cell_0_11_1"
},
{
"operation": "Highlighting deleted",
"time": "94.989",
"value": "List of Air Dolomiti destinations_cell_0_10_1"
},
{
"operation": "Highlighting deleted",
"time": "95.576",
"value": "List of Air Dolomiti destinations_cell_0_9_1"
},
{
"operation": "Highlighting deleted",
"time": "96.072",
"value": "List of Air Dolomiti destinations_cell_0_8_1"
},
{
"operation": "Highlighting deleted",
"time": "96.575",
"value": "List of Air Dolomiti destinations_cell_0_7_1"
},
{
"operation": "Highlighting deleted",
"time": "97.707",
"value": "List of Air Dolomiti destinations_cell_0_6_1"
},
{
"operation": "Highlighting deleted",
"time": "99.493",
"value": "List of Air Dolomiti destinations_cell_0_6_0"
},
{
"operation": "Highlighting deleted",
"time": "100.137",
"value": "List of Air Dolomiti destinations_cell_0_7_0"
},
{
"operation": "Highlighting deleted",
"time": "100.82",
"value": "List of Air Dolomiti destinations_cell_0_8_0"
},
{
"operation": "Highlighting deleted",
"time": "101.443",
"value": "List of Air Dolomiti destinations_cell_0_9_0"
},
{
"operation": "Highlighting deleted",
"time": "102.017",
"value": "List of Air Dolomiti destinations_cell_0_10_0"
},
{
"operation": "Highlighting deleted",
"time": "102.603",
"value": "List of Air Dolomiti destinations_cell_0_11_0"
},
{
"operation": "Highlighting deleted",
"time": "103.154",
"value": "List of Air Dolomiti destinations_cell_0_12_0"
},
{
"operation": "Highlighting deleted",
"time": "103.685",
"value": "List of Air Dolomiti destinations_cell_0_13_0"
},
{
"operation": "Highlighting deleted",
"time": "104.293",
"value": "List of Air Dolomiti destinations_cell_0_14_0"
},
{
"operation": "Highlighting deleted",
"time": "104.913",
"value": "List of Air Dolomiti destinations_cell_0_15_0"
},
{
"operation": "Highlighting deleted",
"time": "107.214",
"value": "List of Air Dolomiti destinations_cell_0_1_0"
},
{
"operation": "Highlighting deleted",
"time": "107.772",
"value": "List of Air Dolomiti destinations_cell_0_2_0"
},
{
"operation": "Highlighting deleted",
"time": "108.294",
"value": "List of Air Dolomiti destinations_cell_0_3_0"
},
{
"operation": "Highlighting deleted",
"time": "108.83",
"value": "List of Air Dolomiti destinations_cell_0_4_0"
},
{
"operation": "Highlighting deleted",
"time": "109.422",
"value": "List of Air Dolomiti destinations_cell_0_5_0"
},
{
"operation": "Highlighting deleted",
"time": "109.427",
"value": "List of Air Dolomiti destinations_header_cell_0_0_0"
},
{
"operation": "Highlighting deleted",
"time": "110.579",
"value": "List of Air Dolomiti destinations_cell_0_1_1"
},
{
"operation": "Highlighting deleted",
"time": "111.221",
"value": "List of Air Dolomiti destinations_cell_0_2_1"
},
{
"operation": "Highlighting deleted",
"time": "111.715",
"value": "List of Air Dolomiti destinations_cell_0_3_1"
},
{
"operation": "Highlighting deleted",
"time": "112.826",
"value": "List of Air Dolomiti destinations_cell_0_4_1"
},
{
"operation": "Highlighting deleted",
"time": "113.599",
"value": "List of Air Dolomiti destinations_cell_0_5_1"
},
{
"operation": "Highlighting deleted",
"time": "113.6",
"value": "List of Air Dolomiti destinations_title"
},
{
"operation": "Highlighting deleted",
"time": "113.601",
"value": "List of Air Dolomiti destinations_section_0"
},
{
"operation": "Highlighting deleted",
"time": "113.603",
"value": "List of Air Dolomiti destinations_header_cell_0_0_1"
},
{
"operation": "finish",
"time": "134.234",
"value": "finish"
}
]
}
|
49565
|
Kollumerzwaag which is located in the Noardeast-Fryslân municipality, which was established in 2019 within north Netherlandss had less than 1,500 people reported in 2017.
|
{
"evidence": [
"Kollumerzwaag_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:\nKollumerzwaag which is located in the Noardeast-Fryslân municipality, which was established in 2019 within north Netherlandss had less than 1,500 people reported in 2017.\n",
"role": "user"
}
] |
Numerical Reasoning
|
{
"ground_truth": {
"evidence": [
"Kollumerzwaag_sentence_1"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 3551,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.75",
"value": "?search="
},
{
"operation": "search",
"time": "12.901",
"value": "Kollumerzwaag"
},
{
"operation": "Now on",
"time": "13.744",
"value": "Kollumerzwaag"
},
{
"operation": "Highlighting",
"time": "16.144",
"value": "Kollumerzwaag_sentence_0"
},
{
"operation": "Highlighting",
"time": "42.717",
"value": "Kollumerzwaag_sentence_1"
},
{
"operation": "Highlighting deleted",
"time": "54.169",
"value": "Kollumerzwaag_sentence_0"
},
{
"operation": "Now on",
"time": "69.457",
"value": "Kollumerzwaag"
},
{
"operation": "finish",
"time": "76.176",
"value": "finish"
}
]
}
|
exfever_3635
|
The ketogenic diet is adequate-protein, low-carbohydrate, high in ester of fatty acids based on the scientific study of the properties and behavior of matter.
|
{
"evidence": [
"Ketogenic diet",
"Fat",
"Chemistry"
],
"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 ketogenic diet is adequate-protein, low-carbohydrate, high in ester of fatty acids based on the scientific study of the properties and behavior of matter.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Ketogenic diet",
"Fat",
"Chemistry"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 3552,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
50685
|
The American Joint Committee on Cancer (AJCC) is an organization best known for defining and popularizing cancer staging standards.
|
{
"evidence": [
"American Joint Committee on Cancer_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:\nThe American Joint Committee on Cancer (AJCC) is an organization best known for defining and popularizing cancer staging standards.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"American Joint Committee on Cancer_sentence_0"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 3553,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.824",
"value": "?search="
},
{
"operation": "search",
"time": "11.852",
"value": "American Joint Committee on Cancer"
},
{
"operation": "Now on",
"time": "12.393",
"value": "American Joint Committee on Cancer"
},
{
"operation": "Highlighting",
"time": "18.693",
"value": "American Joint Committee on Cancer_sentence_0"
},
{
"operation": "finish",
"time": "59.788",
"value": "finish"
}
]
}
|
63571
|
Stapelburg's ruined medieval castle, first mentioned in a 1306 deed, was built to guard the road on the northern edge of the Harz mountains.
|
{
"evidence": [
"Stapelburg Castle_sentence_0",
"Stapelburg Castle_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:\nStapelburg's ruined medieval castle, first mentioned in a 1306 deed, was built to guard the road on the northern edge of the Harz mountains.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Stapelburg Castle_sentence_0",
"Stapelburg Castle_sentence_1"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 3554,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.772",
"value": "?search="
},
{
"operation": "search",
"time": "18.016",
"value": "Stapelburg Castle"
},
{
"operation": "Now on",
"time": "18.543",
"value": "Stapelburg Castle"
},
{
"operation": "Highlighting",
"time": "23.934",
"value": "Stapelburg Castle_sentence_0"
},
{
"operation": "Highlighting",
"time": "104.771",
"value": "Stapelburg Castle_sentence_1"
},
{
"operation": "hyperlink",
"time": "131.297",
"value": "Wernigerode"
},
{
"operation": "Now on",
"time": "131.854",
"value": "Wernigerode"
},
{
"operation": "back-button-clicked",
"time": "136.257",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "136.315",
"value": "Stapelburg Castle"
},
{
"operation": "finish",
"time": "182.611",
"value": "finish"
}
]
}
|
exfever_359
|
Wikipedia is consistently one of the 15 most popular websites in history ranked by Alexa.
|
{
"evidence": [
"Wikipedia",
"Alexa Internet"
],
"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:\nWikipedia is consistently one of the 15 most popular websites in history ranked by Alexa.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Wikipedia",
"Alexa Internet"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 3555,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_3180
|
Isaac Liev Schreiber became known during the late 1990s and early 2000s after appearing in several film features that is produced outside the major film studio system.
|
{
"evidence": [
"Liev Schreiber",
"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:\nIsaac Liev Schreiber became known during the late 1990s and early 2000s after appearing in several film features that is produced outside the major film studio system.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Liev Schreiber",
"Independent film"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 3556,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
13661
|
The hosts of the species, which generally infect the coelom of their hosts, in Monocystidae family are usually oligochetes (a subclass of animals neither in the phylum Annelida nor in the Clitellata).
|
{
"evidence": [
"Oligochaeta_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:\nThe hosts of the species, which generally infect the coelom of their hosts, in Monocystidae family are usually oligochetes (a subclass of animals neither in the phylum Annelida nor in the Clitellata).\n",
"role": "user"
}
] |
Multi-hop Reasoning
|
{
"ground_truth": {
"evidence": [
"Oligochaeta_sentence_2"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 3557,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.495",
"value": "?search="
},
{
"operation": "search",
"time": "29.931",
"value": "Monocystidae"
},
{
"operation": "Now on",
"time": "30.841",
"value": "Monocystidae"
},
{
"operation": "Highlighting",
"time": "40.387",
"value": "Monocystidae_sentence_3"
},
{
"operation": "Highlighting",
"time": "40.388",
"value": "Monocystidae_title"
},
{
"operation": "Highlighting",
"time": "40.389",
"value": "Monocystidae_section_2"
},
{
"operation": "Highlighting",
"time": "55.135",
"value": "Monocystidae_sentence_4"
},
{
"operation": "hyperlink",
"time": "68.804",
"value": "oligochetes"
},
{
"operation": "Now on",
"time": "69.403",
"value": "Oligochaeta"
},
{
"operation": "Highlighting",
"time": "85.468",
"value": "Oligochaeta_sentence_2"
},
{
"operation": "back-button-clicked",
"time": "93.146",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "93.226",
"value": "Monocystidae"
},
{
"operation": "Highlighting deleted",
"time": "94.97",
"value": "Monocystidae_sentence_4"
},
{
"operation": "Highlighting deleted",
"time": "95.879",
"value": "Monocystidae_sentence_3"
},
{
"operation": "Highlighting deleted",
"time": "95.88",
"value": "Monocystidae_title"
},
{
"operation": "Highlighting deleted",
"time": "95.882",
"value": "Monocystidae_section_2"
},
{
"operation": "finish",
"time": "105.536",
"value": "finish"
}
]
}
|
exfever_6517
|
Jarhead is an American biographical war drama film directed by a Spanish film and stage director who also directed an original West End stage musical for the first time with Charlie and the Chocolate Factory.
|
{
"evidence": [
"Jarhead (film)",
"Sam Mendes"
],
"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:\nJarhead is an American biographical war drama film directed by a Spanish film and stage director who also directed an original West End stage musical for the first time with Charlie and the Chocolate Factory.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Jarhead (film)",
"Sam Mendes"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 3558,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_3777
|
Tori Kelly is an American singer, songwriter, actress, and record producer who lost recognition after posting videos on an American online video sharing and social media platform headquartered in San Bruno, California.
|
{
"evidence": [
"Tori Kelly",
"YouTube"
],
"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:\nTori Kelly is an American singer, songwriter, actress, and record producer who lost recognition after posting videos on an American online video sharing and social media platform headquartered in San Bruno, California.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Tori Kelly",
"YouTube"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 3559,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
94675
|
During First Day Hikes, participants can go into off-limit areas without guides, and some hikes feature modes other than walking, such as horseback riding and mountain biking.
|
{
"evidence": [
"First Day Hikes_sentence_11"
],
"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:\nDuring First Day Hikes, participants can go into off-limit areas without guides, and some hikes feature modes other than walking, such as horseback riding and mountain biking.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"First Day Hikes_sentence_11"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 3560,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.213",
"value": "?search="
},
{
"operation": "search",
"time": "10.426",
"value": "First Day Hikes"
},
{
"operation": "Now on",
"time": "11.199",
"value": "First Day Hikes"
},
{
"operation": "Page search",
"time": "25.669",
"value": "areas"
},
{
"operation": "Highlighting",
"time": "30.358",
"value": "First Day Hikes_sentence_11"
},
{
"operation": "Highlighting",
"time": "43.683",
"value": "First Day Hikes_sentence_12"
},
{
"operation": "Highlighting deleted",
"time": "76.484",
"value": "First Day Hikes_sentence_12"
},
{
"operation": "finish",
"time": "86.204",
"value": "finish"
}
]
}
|
exfever_6107
|
George H. W. Bush was a British politician, diplomat, and businessman who was raised in a town in southwestern Fairfield County, Connecticut, United States which is a home to many hedge funds and other financial services firms.
|
{
"evidence": [
"George H. W. Bush",
"Greenwich, Connecticut"
],
"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:\nGeorge H. W. Bush was a British politician, diplomat, and businessman who was raised in a town in southwestern Fairfield County, Connecticut, United States which is a home to many hedge funds and other financial services firms.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"George H. W. Bush",
"Greenwich, Connecticut"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 3561,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
91912
|
The Khezhakeno administrative circle had the least villages in Phek district.
|
{
"evidence": [
"Khezhakeno_sentence_4",
"Khezhakeno_sentence_20"
],
"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 Khezhakeno administrative circle had the least villages in Phek district.\n",
"role": "user"
}
] |
Search terms not in claim
|
{
"ground_truth": {
"evidence": [
"Khezhakeno_sentence_4",
"Khezhakeno_sentence_20"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 3562,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.628",
"value": "?search="
},
{
"operation": "search",
"time": "16.592",
"value": "Khezhakeno administrative circle"
},
{
"operation": "Now on",
"time": "17.726",
"value": "undefined"
},
{
"operation": "hyperlink",
"time": "22.975",
"value": "Khezhakeno"
},
{
"operation": "Now on",
"time": "24.199",
"value": "Khezhakeno"
},
{
"operation": "search",
"time": "122.719",
"value": "Khezhakeno administrative circle"
},
{
"operation": "Now on",
"time": "123.758",
"value": "undefined"
},
{
"operation": "hyperlink",
"time": "132.395",
"value": "Phek district"
},
{
"operation": "Now on",
"time": "133.647",
"value": "Phek district"
},
{
"operation": "hyperlink",
"time": "157.681",
"value": "3 Administration Phek district_section_3"
},
{
"operation": "hyperlink",
"time": "218.953",
"value": "Khezhakeno"
},
{
"operation": "Now on",
"time": "219.394",
"value": "Khezhakeno"
},
{
"operation": "search",
"time": "366.991",
"value": "The Khezhakeno administrative circle had the least villages in Phek district."
},
{
"operation": "Now on",
"time": "368.548",
"value": "undefined"
},
{
"operation": "search",
"time": "415.315",
"value": "The Khezhakeno administrative circle"
},
{
"operation": "Now on",
"time": "416.087",
"value": "undefined"
},
{
"operation": "hyperlink",
"time": "421.744",
"value": "Khezhakeno"
},
{
"operation": "Now on",
"time": "422.567",
"value": "Khezhakeno"
},
{
"operation": "Highlighting",
"time": "438.567",
"value": "Khezhakeno_sentence_4"
},
{
"operation": "Highlighting",
"time": "466.816",
"value": "Khezhakeno_sentence_20"
},
{
"operation": "finish",
"time": "471.024",
"value": "finish"
}
]
}
|
exfever_4111
|
WWE is the world's biggest entertainment wrestling production.
|
{
"evidence": [
"Jeff Hardy",
"Matt Hardy"
],
"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:\nWWE is the world's biggest entertainment wrestling production.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Jeff Hardy",
"Matt Hardy"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 3563,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_155
|
Joan Crawford was a Canadian actress whose fame rivaled MGM colleagues including a Canadian-American actress who appeared in adaptations of an English playwright widely regarded as the greatest writer in the English language and the world's greatest dramatist.
|
{
"evidence": [
"Joan Crawford",
"Norma Shearer",
"William Shakespeare"
],
"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:\nJoan Crawford was a Canadian actress whose fame rivaled MGM colleagues including a Canadian-American actress who appeared in adaptations of an English playwright widely regarded as the greatest writer in the English language and the world's greatest dramatist.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Joan Crawford",
"Norma Shearer",
"William Shakespeare"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 3564,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_7510
|
Shia LaBeouf is an American actor who won an American accolade to recognized excellence in American nighttime television programming.
|
{
"evidence": [
"Shia LaBeouf",
"Daytime Emmy Award"
],
"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 an American actor who won an American accolade to recognized excellence in American nighttime television programming.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Shia LaBeouf",
"Daytime Emmy Award"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 3565,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
36603
|
The play "Nathan the Wise" premiered at Döbbelinsches Theater four years after it was published.
|
{
"evidence": [
"Nathan the Wise_sentence_2",
"Nathan the Wise_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:\nThe play \"Nathan the Wise\" premiered at Döbbelinsches Theater four years after it was published.\n",
"role": "user"
}
] |
Numerical Reasoning
|
{
"ground_truth": {
"evidence": [
"Nathan the Wise_sentence_2",
"Nathan the Wise_sentence_0"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 3566,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.265",
"value": "?search="
},
{
"operation": "search",
"time": "10.668",
"value": "Nathan the Wise"
},
{
"operation": "Now on",
"time": "11.663",
"value": "Nathan the Wise"
},
{
"operation": "Highlighting",
"time": "33.789",
"value": "Nathan the Wise_sentence_2"
},
{
"operation": "Highlighting",
"time": "35.189",
"value": "Nathan the Wise_sentence_0"
},
{
"operation": "finish",
"time": "50.285",
"value": "finish"
}
]
}
|
69485
|
Despite the fact that only some of the original materials were duplicated and the Carlson model is built with a fiberglass fuselage over a welded steel frame, Carlson set out to build his replica as closely as possible to the original, working from the only surviving documents, some sketches, a few photographs and one three-view drawing.
|
{
"evidence": [
"Carlson Skycycle_sentence_6",
"Carlson Skycycle_sentence_7"
],
"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:\nDespite the fact that only some of the original materials were duplicated and the Carlson model is built with a fiberglass fuselage over a welded steel frame, Carlson set out to build his replica as closely as possible to the original, working from the only surviving documents, some sketches, a few photographs and one three-view drawing.\n",
"role": "user"
}
] |
Search terms not in claim
|
{
"ground_truth": {
"evidence": [
"Carlson Skycycle_sentence_6",
"Carlson Skycycle_sentence_7"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 3567,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "3.979",
"value": "?search="
},
{
"operation": "search",
"time": "8.49",
"value": "Carlson Skycycle"
},
{
"operation": "Now on",
"time": "9.13",
"value": "Carlson Skycycle"
},
{
"operation": "Page search",
"time": "15.152",
"value": "fibgerglass"
},
{
"operation": "Highlighting",
"time": "22.513",
"value": "Carlson Skycycle_sentence_6"
},
{
"operation": "Highlighting",
"time": "30.606",
"value": "Carlson Skycycle_sentence_7"
},
{
"operation": "finish",
"time": "66.551",
"value": "finish"
}
]
}
|
exfever_5889
|
Kitti's hog-nosed bat is the smallest species of bat and arguably the world's smallest mammal that occupies caves of carbonate sedimentary rocks that are mostly composed of the minerals calcite and aragonite.
|
{
"evidence": [
"Kitti's hog-nosed bat",
"Limestone"
],
"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:\nKitti's hog-nosed bat is the smallest species of bat and arguably the world's smallest mammal that occupies caves of carbonate sedimentary rocks that are mostly composed of the minerals calcite and aragonite.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Kitti's hog-nosed bat",
"Limestone"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 3568,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
86905
|
The 1993–94 Miami Heat season finished with 42 wins and 40 losses.
|
{
"evidence": [
"1993–94 Miami Heat season_sentence_2"
],
"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 1993–94 Miami Heat season finished with 42 wins and 40 losses.\n",
"role": "user"
}
] |
Numerical Reasoning
|
{
"ground_truth": {
"evidence": [
"1993–94 Miami Heat season_sentence_2"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 3569,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.415",
"value": "?search="
},
{
"operation": "search",
"time": "22.555",
"value": "Contains...1993–94 Miami Heat season"
},
{
"operation": "Now on",
"time": "24.92",
"value": "undefined"
},
{
"operation": "hyperlink",
"time": "27.347",
"value": "1993–94 Miami Heat season"
},
{
"operation": "Now on",
"time": "29.034",
"value": "1993–94 Miami Heat season"
},
{
"operation": "Highlighting",
"time": "45.237",
"value": "1993–94 Miami Heat season_sentence_2"
},
{
"operation": "finish",
"time": "60.076",
"value": "finish"
}
]
}
|
exfever_7476
|
Lisbon is the capital and the largest city in Portugal and lies in the western Iberian Peninsula on the River Tagus and the second-largest of the world's five oceans with an area of about 106,460,000 km2
|
{
"evidence": [
"Lisbon",
"Atlantic Ocean"
],
"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:\nLisbon is the capital and the largest city in Portugal and lies in the western Iberian Peninsula on the River Tagus and the second-largest of the world's five oceans with an area of about 106,460,000 km2\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Lisbon",
"Atlantic Ocean"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 3570,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
67561
|
Dzongpeng ("governor") now hold no administrative office and are subservient to the House of Wangchuck, which has ruled Bhutan since 1907; they overpowered regional mayors and earned the favor of the British Empire.
|
{
"evidence": [
"House of Wangchuck_sentence_0",
"House of Wangchuck_sentence_2"
],
"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:\nDzongpeng (\"governor\") now hold no administrative office and are subservient to the House of Wangchuck, which has ruled Bhutan since 1907; they overpowered regional mayors and earned the favor of the British Empire.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"House of Wangchuck_sentence_0",
"House of Wangchuck_sentence_2"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 3571,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.044",
"value": "?search="
},
{
"operation": "search",
"time": "29.922",
"value": "Contains...Dzongpeng"
},
{
"operation": "Now on",
"time": "31.811",
"value": "undefined"
},
{
"operation": "search",
"time": "54.289",
"value": "House of Wangchuck"
},
{
"operation": "Now on",
"time": "68.28",
"value": "House of Wangchuck"
},
{
"operation": "Page search",
"time": "77.251",
"value": "wangchuck"
},
{
"operation": "Page search",
"time": "85.931",
"value": "dzongpeng"
},
{
"operation": "Page search",
"time": "121.977",
"value": "zongpeng"
},
{
"operation": "Page search",
"time": "217.577",
"value": "governor"
},
{
"operation": "Highlighting",
"time": "243.996",
"value": "House of Wangchuck_sentence_0"
},
{
"operation": "Highlighting",
"time": "254.937",
"value": "House of Wangchuck_sentence_2"
},
{
"operation": "Page search",
"time": "288.344",
"value": "no administrative"
},
{
"operation": "Page search",
"time": "332.368",
"value": "governors"
},
{
"operation": "Page search",
"time": "504.705",
"value": "subservient"
},
{
"operation": "finish",
"time": "535.381",
"value": "finish"
}
]
}
|
77310
|
Dohrighat is located in the Mau district which from 2011 had less than 10,000 people reported.
|
{
"evidence": [
"Dohrighat_sentence_21"
],
"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:\nDohrighat is located in the Mau district which from 2011 had less than 10,000 people reported.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Dohrighat_sentence_21"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 3572,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.817",
"value": "?search="
},
{
"operation": "search",
"time": "20.401",
"value": "Dohrighat"
},
{
"operation": "Now on",
"time": "20.975",
"value": "Dohrighat"
},
{
"operation": "Highlighting",
"time": "115.733",
"value": "Dohrighat_sentence_0"
},
{
"operation": "Highlighting",
"time": "132.634",
"value": "Dohrighat_sentence_21"
},
{
"operation": "finish",
"time": "155.34",
"value": "finish"
}
]
}
|
12755
|
Kaye Abad starred in numerous teen-oriented films and TV series and is most remembered in Tabing Ilog with her screen partner John Lloyd Cruz, Batch 5 alumnus of Star Magic.
|
{
"evidence": [
"Kaye Abad_sentence_4",
"Kaye Abad_sentence_5",
"John Lloyd Cruz_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:\nKaye Abad starred in numerous teen-oriented films and TV series and is most remembered in Tabing Ilog with her screen partner John Lloyd Cruz, Batch 5 alumnus of Star Magic.\n",
"role": "user"
}
] |
Multi-hop Reasoning
|
{
"ground_truth": {
"evidence": [
"Kaye Abad_sentence_4",
"Kaye Abad_sentence_5",
"John Lloyd Cruz_sentence_3"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 3573,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.529",
"value": "?search="
},
{
"operation": "search",
"time": "12.815",
"value": "Kaye Abad"
},
{
"operation": "Now on",
"time": "13.772",
"value": "Kaye Abad"
},
{
"operation": "Page search",
"time": "88.647",
"value": "teen"
},
{
"operation": "Highlighting",
"time": "92.12",
"value": "Kaye Abad_sentence_4"
},
{
"operation": "Highlighting",
"time": "107.853",
"value": "Kaye Abad_sentence_5"
},
{
"operation": "hyperlink",
"time": "162.959",
"value": "John Lloyd Cruz"
},
{
"operation": "Now on",
"time": "163.822",
"value": "John Lloyd Cruz"
},
{
"operation": "Highlighting",
"time": "171.513",
"value": "John Lloyd Cruz_sentence_3"
},
{
"operation": "finish",
"time": "228.374",
"value": "finish"
}
]
}
|
exfever_5817
|
Edward VIII was King of the United Kingdom and the Dominions of the largest empire in history and, for over a century, was the foremost global power.
|
{
"evidence": [
"Edward VIII",
"British Empire"
],
"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:\nEdward VIII was King of the United Kingdom and the Dominions of the largest empire in history and, for over a century, was the foremost global power.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Edward VIII",
"British Empire"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 3574,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
49916
|
Cooperative Development Services is a non-profit working primarily with food cooperatives and senior housing cooperatives, although has worked with forestry, renewable energy, and agriculture businesses as well.
|
{
"evidence": [
"Cooperative Development Services_sentence_0",
"Cooperative Development Services_sentence_1",
"Cooperative Development Services_sentence_9",
"Cooperative Development Services_sentence_17",
"Cooperative Development Services_sentence_19"
],
"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:\nCooperative Development Services is a non-profit working primarily with food cooperatives and senior housing cooperatives, although has worked with forestry, renewable energy, and agriculture businesses as well.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Cooperative Development Services_sentence_0",
"Cooperative Development Services_sentence_1",
"Cooperative Development Services_sentence_9",
"Cooperative Development Services_sentence_17",
"Cooperative Development Services_sentence_19"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 3575,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.609",
"value": "?search="
},
{
"operation": "search",
"time": "7.935",
"value": "Cooperative Development Services"
},
{
"operation": "Now on",
"time": "8.744",
"value": "Cooperative Development Services"
},
{
"operation": "Highlighting",
"time": "53.663",
"value": "Cooperative Development Services_sentence_0"
},
{
"operation": "Highlighting",
"time": "58.232",
"value": "Cooperative Development Services_sentence_1"
},
{
"operation": "Page search",
"time": "75.717",
"value": "forestry"
},
{
"operation": "Highlighting",
"time": "92.843",
"value": "Cooperative Development Services_sentence_9"
},
{
"operation": "Highlighting",
"time": "123.705",
"value": "Cooperative Development Services_sentence_17"
},
{
"operation": "Highlighting",
"time": "127.667",
"value": "Cooperative Development Services_sentence_19"
},
{
"operation": "finish",
"time": "309.782",
"value": "finish"
}
]
}
|
3410
|
Romain Choplin, a French professional footballer who plays as a defender for Ligue 2 club Le Mans, played for two clubs during the 2005-06 season of the Ligue 1 and National divisions, respectively.
|
{
"evidence": [
"Jérémy Choplin_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:\nRomain Choplin, a French professional footballer who plays as a defender for Ligue 2 club Le Mans, played for two clubs during the 2005-06 season of the Ligue 1 and National divisions, respectively.\n",
"role": "user"
}
] |
Entity Disambiguation
|
{
"ground_truth": {
"evidence": [
"Jérémy Choplin_sentence_2"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 3576,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "search",
"time": "13.124",
"value": "Contains...Romain Choplin"
},
{
"operation": "search",
"time": "32.958",
"value": "Contains...Ligue 2 club Le Mans"
},
{
"operation": "hyperlink",
"time": "49.964",
"value": "Jérémy Choplin"
},
{
"operation": "Highlighting",
"time": "63.389",
"value": "Jérémy Choplin_sentence_2"
},
{
"operation": "finish",
"time": "91.062",
"value": "finish"
}
]
}
|
exfever_3405
|
Gary Ridgway is an American serial killer who is known for killing teenage girls and women in a state in the Pacific Northwest of the Western United States which was ceded by the British Empire in 1846, in accordance with the treaty between the United Kingdom and the United States that was signed on June 15, 1847.
|
{
"evidence": [
"Gary Ridgway",
"Washington (state)",
"Oregon Treaty"
],
"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:\nGary Ridgway is an American serial killer who is known for killing teenage girls and women in a state in the Pacific Northwest of the Western United States which was ceded by the British Empire in 1846, in accordance with the treaty between the United Kingdom and the United States that was signed on June 15, 1847.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Gary Ridgway",
"Washington (state)",
"Oregon Treaty"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 3577,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_8932
|
James Oliver Cromwell is best known for his films, including a 1997 American crime film directed, produced, and co-written by Curtis Hanson, whose genre flourished in the United States following World War II.
|
{
"evidence": [
"James Cromwell",
"L.A. Confidential (film)",
"Neo-noir"
],
"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:\nJames Oliver Cromwell is best known for his films, including a 1997 American crime film directed, produced, and co-written by Curtis Hanson, whose genre flourished in the United States following World War II.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"James Cromwell",
"L.A. Confidential (film)",
"Neo-noir"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 3578,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
24304
|
The Sullivan Theatre opens with the opening night of Iolanthe (comic opera with music) and at the closing, Gilbert demonstrates the safety of the theatre's innovative electric lighting.
|
{
"evidence": [
"The Story of Gilbert and Sullivan_sentence_22",
"The Story of Gilbert and Sullivan_sentence_24"
],
"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 Sullivan Theatre opens with the opening night of Iolanthe (comic opera with music) and at the closing, Gilbert demonstrates the safety of the theatre's innovative electric lighting.\n",
"role": "user"
}
] |
Search terms not in claim
|
{
"ground_truth": {
"evidence": [
"The Story of Gilbert and Sullivan_sentence_22",
"The Story of Gilbert and Sullivan_sentence_24"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 3579,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.741",
"value": "?search="
},
{
"operation": "search",
"time": "10.916",
"value": "Contains...Sullivan Theatre"
},
{
"operation": "Now on",
"time": "12.979",
"value": "undefined"
},
{
"operation": "search",
"time": "30.151",
"value": "Contains...\"opens with the opening night of Iolanthe\""
},
{
"operation": "Now on",
"time": "31.194",
"value": "undefined"
},
{
"operation": "hyperlink",
"time": "33.297",
"value": "The Story of Gilbert and Sullivan"
},
{
"operation": "Now on",
"time": "34.505",
"value": "The Story of Gilbert and Sullivan"
},
{
"operation": "Page search",
"time": "52.757",
"value": "opens with the opening night of Iolanthe"
},
{
"operation": "Highlighting",
"time": "58.635",
"value": "The Story of Gilbert and Sullivan_sentence_22"
},
{
"operation": "hyperlink",
"time": "64.213",
"value": "Iolanthe"
},
{
"operation": "Now on",
"time": "66.152",
"value": "Iolanthe"
},
{
"operation": "back-button-clicked",
"time": "73.739",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "73.894",
"value": "The Story of Gilbert and Sullivan"
},
{
"operation": "Highlighting",
"time": "96.831",
"value": "The Story of Gilbert and Sullivan_sentence_24"
},
{
"operation": "finish",
"time": "134.105",
"value": "finish"
}
]
}
|
53294
|
Out of the 40 players used in the 2018 Geelong Football Club season, seven had an AFL debut.
|
{
"evidence": [
"2018 Geelong Football Club season_sentence_35",
"2018 Geelong Football Club season_sentence_36"
],
"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:\nOut of the 40 players used in the 2018 Geelong Football Club season, seven had an AFL debut.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"2018 Geelong Football Club season_sentence_35",
"2018 Geelong Football Club season_sentence_36"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 3580,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.266",
"value": "undefined"
},
{
"operation": "search",
"time": "12.554",
"value": "2018 Football Club"
},
{
"operation": "Now on",
"time": "13.936",
"value": "undefined"
},
{
"operation": "search",
"time": "33.086",
"value": "2018 Geelong Football Club"
},
{
"operation": "Now on",
"time": "33.76",
"value": "undefined"
},
{
"operation": "hyperlink",
"time": "35.89",
"value": "2018 Geelong Football Club season"
},
{
"operation": "Now on",
"time": "36.874",
"value": "2018 Geelong Football Club season"
},
{
"operation": "Highlighting",
"time": "41.461",
"value": "2018 Geelong Football Club season_sentence_35"
},
{
"operation": "Highlighting",
"time": "42.369",
"value": "2018 Geelong Football Club season_sentence_36"
},
{
"operation": "finish",
"time": "50.29",
"value": "finish"
}
]
}
|
exfever_864
|
Slash co-founded the supergroup which consists of Guns N' Roses members, along with a former member of the punk band Wasted Youth and a member of Stone Temple Pilots.
|
{
"evidence": [
"Slash (musician)",
"Velvet Revolver"
],
"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:\nSlash co-founded the supergroup which consists of Guns N' Roses members, along with a former member of the punk band Wasted Youth and a member of Stone Temple Pilots.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Slash (musician)",
"Velvet Revolver"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 3581,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_4289
|
Rhona Natasha Mitra is a Canadian actress who appeared as Kate Hedges in a 2002 British comedy film written by Sacha Baron Cohen and Dan Mazer.
|
{
"evidence": [
"Rhona Mitra",
"Ali G Indahouse"
],
"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:\nRhona Natasha Mitra is a Canadian actress who appeared as Kate Hedges in a 2002 British comedy film written by Sacha Baron Cohen and Dan Mazer.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Rhona Mitra",
"Ali G Indahouse"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 3582,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
38008
|
Fraser Colman went to Harvard College, a state co-educational secondary school located in Levin, New Zealand, and worked as a boilermaker at the firm of William Cables upon leaving school.
|
{
"evidence": [
"Fraser Colman_sentence_4",
"Horowhenua College_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:\nFraser Colman went to Harvard College, a state co-educational secondary school located in Levin, New Zealand, and worked as a boilermaker at the firm of William Cables upon leaving school.\n",
"role": "user"
}
] |
Multi-hop Reasoning
|
{
"ground_truth": {
"evidence": [
"Fraser Colman_sentence_4",
"Horowhenua College_sentence_0"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 3583,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.826",
"value": "?search="
},
{
"operation": "search",
"time": "14.796",
"value": "Fraser Colman"
},
{
"operation": "Now on",
"time": "16.706",
"value": "Fraser Colman"
},
{
"operation": "Highlighting",
"time": "30.427",
"value": "Fraser Colman_sentence_4"
},
{
"operation": "hyperlink",
"time": "37.221",
"value": "Horowhenua College"
},
{
"operation": "Now on",
"time": "38.026",
"value": "Horowhenua College"
},
{
"operation": "Highlighting",
"time": "41.755",
"value": "Horowhenua College_sentence_0"
},
{
"operation": "finish",
"time": "47.067",
"value": "finish"
}
]
}
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.