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_8584
|
Troy Baker is an American musician known for his video game roles in a franchise of first-person shooter games published by Ubisoft for the group of several proprietary graphical operating system families that are developed and marketed by Microsoft.
|
{
"evidence": [
"Troy Baker",
"Far Cry",
"Microsoft Windows"
],
"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:\nTroy Baker is an American musician known for his video game roles in a franchise of first-person shooter games published by Ubisoft for the group of several proprietary graphical operating system families that are developed and marketed by Microsoft.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Troy Baker",
"Far Cry",
"Microsoft Windows"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 4984,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
48344
|
Erik Zabel has never competed in a Milan–San Remo race.
|
{
"evidence": [
"Erik Zabel_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:\nErik Zabel has never competed in a Milan–San Remo race.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Erik Zabel_sentence_3"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 4985,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.895",
"value": "?search="
},
{
"operation": "search",
"time": "7.913",
"value": "Erik Zabel"
},
{
"operation": "Now on",
"time": "9.648",
"value": "Erik Zabel"
},
{
"operation": "Page search",
"time": "47.96",
"value": "milan"
},
{
"operation": "Highlighting",
"time": "51.175",
"value": "Erik Zabel_sentence_3"
},
{
"operation": "finish",
"time": "56.719",
"value": "finish"
}
]
}
|
61738
|
The Port of Spain General Hospital ( once a Colonial hospital that began construction in 1854, using local stone and imported brick facing to provide a Classical style facade) is located in Upper Charlotte Street, Port of Spain, Trinidad and Tobago and opened in 1900.
|
{
"evidence": [
"Port of Spain General Hospital_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 Port of Spain General Hospital ( once a Colonial hospital that began construction in 1854, using local stone and imported brick facing to provide a Classical style facade) is located in Upper Charlotte Street, Port of Spain, Trinidad and Tobago and opened in 1900.\n",
"role": "user"
}
] |
Multi-hop Reasoning
|
{
"ground_truth": {
"evidence": [
"Port of Spain General Hospital_sentence_2"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 4986,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.324",
"value": "?search="
},
{
"operation": "search",
"time": "14.853",
"value": "Port of Spain General Hospital"
},
{
"operation": "Now on",
"time": "15.734",
"value": "Port of Spain General Hospital"
},
{
"operation": "Highlighting",
"time": "21.951",
"value": "Port of Spain General Hospital_sentence_1"
},
{
"operation": "Highlighting",
"time": "34.243",
"value": "Port of Spain General Hospital_sentence_4"
},
{
"operation": "Highlighting deleted",
"time": "44.846",
"value": "Port of Spain General Hospital_sentence_4"
},
{
"operation": "Highlighting deleted",
"time": "45.875",
"value": "Port of Spain General Hospital_sentence_1"
},
{
"operation": "Highlighting",
"time": "49.167",
"value": "Port of Spain General Hospital_sentence_2"
},
{
"operation": "finish",
"time": "65.358",
"value": "finish"
}
]
}
|
exfever_8380
|
Moscovium was first synthesized in 2004 by a joint team of Russian and American scientists at an international research center for nuclear sciences with 5,500 staff members in Dubna, Russia.
|
{
"evidence": [
"Moscovium",
"Joint Institute for Nuclear Research"
],
"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:\nMoscovium was first synthesized in 2004 by a joint team of Russian and American scientists at an international research center for nuclear sciences with 5,500 staff members in Dubna, Russia.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Moscovium",
"Joint Institute for Nuclear Research"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 4987,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
16744
|
MAP2K6 is an enzyme and a member of the dual specificity protein kinase family that is encoded by the MAP2K6 gene on chromosome 17, act as an integration point for multiple biochemical signals and mediates apoptotic cell death in thymocytes.
|
{
"evidence": [
"MAP2K6_sentence_1",
"MAP2K6_sentence_2",
"MAP2K6_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:\nMAP2K6 is an enzyme and a member of the dual specificity protein kinase family that is encoded by the MAP2K6 gene on chromosome 17, act as an integration point for multiple biochemical signals and mediates apoptotic cell death in thymocytes.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"MAP2K6_sentence_1",
"MAP2K6_sentence_2",
"MAP2K6_sentence_3"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 4988,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.753",
"value": "?search="
},
{
"operation": "search",
"time": "14.018",
"value": "MAP2K6"
},
{
"operation": "Now on",
"time": "14.641",
"value": "MAP2K6"
},
{
"operation": "Highlighting",
"time": "90.293",
"value": "MAP2K6_sentence_1"
},
{
"operation": "Highlighting",
"time": "94.773",
"value": "MAP2K6_sentence_2"
},
{
"operation": "Highlighting",
"time": "116.181",
"value": "MAP2K6_sentence_3"
},
{
"operation": "Page search",
"time": "158.323",
"value": "thymocyte"
},
{
"operation": "hyperlink",
"time": "175.26",
"value": "chromosome 17"
},
{
"operation": "Now on",
"time": "176.595",
"value": "Chromosome 17"
},
{
"operation": "Page search",
"time": "181.074",
"value": "thymocyte"
},
{
"operation": "back-button-clicked",
"time": "187.727",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "187.922",
"value": "MAP2K6"
},
{
"operation": "hyperlink",
"time": "196.093",
"value": "gene"
},
{
"operation": "Now on",
"time": "197.426",
"value": "Gene"
},
{
"operation": "back-button-clicked",
"time": "200.999",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "201.215",
"value": "MAP2K6"
},
{
"operation": "hyperlink",
"time": "303.236",
"value": "mitogen-activated protein"
},
{
"operation": "Now on",
"time": "304.757",
"value": "Mitogen-activated protein kinase"
},
{
"operation": "Page search",
"time": "314.691",
"value": "thymocyte"
},
{
"operation": "finish",
"time": "325.833",
"value": "finish"
}
]
}
|
exfever_4778
|
The Armenian genocide was the systematic destruction of the Armenian people in the empire that controlled much of Southeast Europe.
|
{
"evidence": [
"Armenian Genocide",
"Ottoman 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:\nThe Armenian genocide was the systematic destruction of the Armenian people in the empire that controlled much of Southeast Europe.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Armenian Genocide",
"Ottoman Empire"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 4989,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
93337
|
Mercey, first selected for England U19 duty a year ahead of many of his peers, was again called up for England U19 duty in the 2006 World Championships.
|
{
"evidence": [
"Tom Mercey_sentence_13",
"Tom Mercey_sentence_14"
],
"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:\nMercey, first selected for England U19 duty a year ahead of many of his peers, was again called up for England U19 duty in the 2006 World Championships.\n",
"role": "user"
}
] |
Entity Disambiguation
|
{
"ground_truth": {
"evidence": [
"Tom Mercey_sentence_13",
"Tom Mercey_sentence_14"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 4990,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.877",
"value": "?search="
},
{
"operation": "search",
"time": "15.77",
"value": "Mercey"
},
{
"operation": "Now on",
"time": "16.474",
"value": "Mercey"
},
{
"operation": "hyperlink",
"time": "29.577",
"value": "Tom Mercey"
},
{
"operation": "Now on",
"time": "30.163",
"value": "Tom Mercey"
},
{
"operation": "Highlighting",
"time": "48.521",
"value": "Tom Mercey_sentence_13"
},
{
"operation": "Highlighting",
"time": "53.953",
"value": "Tom Mercey_sentence_14"
},
{
"operation": "Highlighting",
"time": "119.335",
"value": "Tom Mercey_sentence_0"
},
{
"operation": "Highlighting deleted",
"time": "119.799",
"value": "Tom Mercey_sentence_0"
},
{
"operation": "finish",
"time": "132.447",
"value": "finish"
}
]
}
|
7590
|
Zack Clayton was inducted into the Naismith Memorial Basketball Hall of Fame in 2017 and he was also a Negro league baseball player and a professional boxing referee.
|
{
"evidence": [
"Zack Clayton_sentence_4",
"Zack Clayton_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:\nZack Clayton was inducted into the Naismith Memorial Basketball Hall of Fame in 2017 and he was also a Negro league baseball player and a professional boxing referee.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Zack Clayton_sentence_4",
"Zack Clayton_sentence_3"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 4991,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.5",
"value": "?search="
},
{
"operation": "search",
"time": "10.398",
"value": "Naismith Memorial Basketball Hall of Fame"
},
{
"operation": "Now on",
"time": "11.158",
"value": "Naismith Memorial Basketball Hall of Fame"
},
{
"operation": "Page search",
"time": "15.25",
"value": "clayton"
},
{
"operation": "hyperlink",
"time": "23.931",
"value": "3 Inductees Naismith Memorial Basketball Hall of Fame_section_3"
},
{
"operation": "hyperlink",
"time": "25.983",
"value": "List of members of the Naismith Memorial Basketball Hall of Fame"
},
{
"operation": "Now on",
"time": "26.783",
"value": "List of members of the Naismith Memorial Basketball Hall of Fame"
},
{
"operation": "Page search",
"time": "30.498",
"value": "clayton"
},
{
"operation": "Page search",
"time": "48.89",
"value": "zack clayton"
},
{
"operation": "search",
"time": "55.258",
"value": "zack clayton"
},
{
"operation": "Now on",
"time": "56.364",
"value": "Zack Clayton"
},
{
"operation": "Page search",
"time": "67.249",
"value": "fame"
},
{
"operation": "Highlighting",
"time": "81.703",
"value": "Zack Clayton_sentence_4"
},
{
"operation": "Highlighting",
"time": "82.469",
"value": "Zack Clayton_sentence_3"
},
{
"operation": "finish",
"time": "109.006",
"value": "finish"
}
]
}
|
76317
|
James Ambrose Cutting was an inventor but did not hold any patents.
|
{
"evidence": [
"James Ambrose Cutting_sentence_10"
],
"label": "REFUTE"
}
|
feverous
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nJames Ambrose Cutting was an inventor but did not hold any patents.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"James Ambrose Cutting_sentence_10"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 4992,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.213",
"value": "?search="
},
{
"operation": "search",
"time": "16.183",
"value": "James Ambrose Cutting"
},
{
"operation": "Now on",
"time": "17.105",
"value": "James Ambrose Cutting"
},
{
"operation": "Highlighting",
"time": "36.039",
"value": "James Ambrose Cutting_sentence_10"
},
{
"operation": "finish",
"time": "45.834",
"value": "finish"
}
]
}
|
exfever_5077
|
Anonymous is a decentralized international group that uses computer-based techniques such as hacking as a form of civil disobedience to promote a political agenda or social change.
|
{
"evidence": [
"Anonymous (group)",
"Hacktivism"
],
"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:\nAnonymous is a decentralized international group that uses computer-based techniques such as hacking as a form of civil disobedience to promote a political agenda or social change.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Anonymous (group)",
"Hacktivism"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 4993,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
6993
|
The Dragon Dentist is an 5- minute long original net animation written and directed by Ōtarō Maijō, voice acted and led by Fumika Shimizu and Xanthe Huynh as Nonoko Kishii and Nobuhiko Okamoto and Xander Mobus as Bell and run from February 18-25 2017
|
{
"evidence": [
"The Dragon Dentist_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:\nThe Dragon Dentist is an 5- minute long original net animation written and directed by Ōtarō Maijō, voice acted and led by Fumika Shimizu and Xanthe Huynh as Nonoko Kishii and Nobuhiko Okamoto and Xander Mobus as Bell and run from February 18-25 2017\n",
"role": "user"
}
] |
Numerical Reasoning
|
{
"ground_truth": {
"evidence": [
"The Dragon Dentist_sentence_0"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 4994,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.219",
"value": "?search="
},
{
"operation": "search",
"time": "16.486",
"value": "Dragon Dentist"
},
{
"operation": "Now on",
"time": "17.51",
"value": "The Dragon Dentist"
},
{
"operation": "Highlighting",
"time": "20.682",
"value": "The Dragon Dentist_sentence_0"
},
{
"operation": "finish",
"time": "47.436",
"value": "finish"
}
]
}
|
47707
|
During the 1947 Boston Braves season, which was the franchise's 77th season, Johnny Said threw the first pitch against against Jackie Robinson in a game against the Brooklyn Dodgers where the final score was 5-3 for Brooklyn.
|
{
"evidence": [
"1947 Boston Braves season_sentence_0",
"1947 Boston Braves season_sentence_2",
"1947 Boston Braves season_sentence_3",
"1947 Boston Braves season_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:\nDuring the 1947 Boston Braves season, which was the franchise's 77th season, Johnny Said threw the first pitch against against Jackie Robinson in a game against the Brooklyn Dodgers where the final score was 5-3 for Brooklyn.\n",
"role": "user"
}
] |
Numerical Reasoning
|
{
"ground_truth": {
"evidence": [
"1947 Boston Braves season_sentence_0",
"1947 Boston Braves season_sentence_2",
"1947 Boston Braves season_sentence_3",
"1947 Boston Braves season_sentence_5"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 4995,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.792",
"value": "?search="
},
{
"operation": "search",
"time": "15.838",
"value": "1947 Boston Braves season"
},
{
"operation": "Now on",
"time": "17.237",
"value": "1947 Boston Braves season"
},
{
"operation": "Highlighting",
"time": "36.877",
"value": "1947 Boston Braves season_sentence_0"
},
{
"operation": "Highlighting",
"time": "44.716",
"value": "1947 Boston Braves season_sentence_2"
},
{
"operation": "Highlighting",
"time": "45.538",
"value": "1947 Boston Braves season_sentence_3"
},
{
"operation": "Highlighting",
"time": "54.479",
"value": "1947 Boston Braves season_sentence_5"
},
{
"operation": "finish",
"time": "70.937",
"value": "finish"
}
]
}
|
exfever_46
|
James VI and I was King of England and Ireland as James I from the accession that followed the death of James's cousin, the Queen of England and Ireland, from 17 November 1558 until 1604.
|
{
"evidence": [
"James VI and I",
"Union of the Crowns",
"Elizabeth I of England"
],
"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:\nJames VI and I was King of England and Ireland as James I from the accession that followed the death of James's cousin, the Queen of England and Ireland, from 17 November 1558 until 1604.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"James VI and I",
"Union of the Crowns",
"Elizabeth I of England"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 4996,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_6117
|
Johnny Mnemonic is a 1995 cyberpunk film that portrays the view of the future in the world dominated by a corporation that is a massive conglomerate holding monopolistic or near-monopolistic control over multiple markets.
|
{
"evidence": [
"Johnny Mnemonic (film)",
"Megacorporation"
],
"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:\nJohnny Mnemonic is a 1995 cyberpunk film that portrays the view of the future in the world dominated by a corporation that is a massive conglomerate holding monopolistic or near-monopolistic control over multiple markets.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Johnny Mnemonic (film)",
"Megacorporation"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 4997,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_4708
|
Luxo Jr. is a 1986 American computer-animated motion picture that is short enough in running time not to be considered a feature film produced and released by Pixar in 1986.
|
{
"evidence": [
"Luxo Jr.",
"Short 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:\nLuxo Jr. is a 1986 American computer-animated motion picture that is short enough in running time not to be considered a feature film produced and released by Pixar in 1986.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Luxo Jr.",
"Short film"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 4998,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_2377
|
Elisabeth Sladen was an English actress best known in the television series Doctor Who as the investigative journalist dedicated to protecting Venus from alien invasions.
|
{
"evidence": [
"Elisabeth Sladen",
"Sarah Jane Smith"
],
"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:\nElisabeth Sladen was an English actress best known in the television series Doctor Who as the investigative journalist dedicated to protecting Venus from alien invasions.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Elisabeth Sladen",
"Sarah Jane Smith"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 4999,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_99
|
Gordon Ramsay played football and chose to play for Warwickshire.
|
{
"evidence": [
"Hotel Hell",
"Ramsay's Kitchen Nightmares",
"British Academy of Film and Television Arts"
],
"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:\nGordon Ramsay played football and chose to play for Warwickshire.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Hotel Hell",
"Ramsay's Kitchen Nightmares",
"British Academy of Film and Television Arts"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 5000,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
24261
|
Isan people overran the Lao Wang and forced them into corvee (unpaid intermittent labour expected by landowners or monarchs) labour before they were integrated into Isan society; much of Isan was populated this way.
|
{
"evidence": [
"Isan people_sentence_38"
],
"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:\nIsan people overran the Lao Wang and forced them into corvee (unpaid intermittent labour expected by landowners or monarchs) labour before they were integrated into Isan society; much of Isan was populated this way.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Isan people_sentence_38"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 5001,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.566",
"value": "?search="
},
{
"operation": "search",
"time": "27.593",
"value": "Isan"
},
{
"operation": "Now on",
"time": "28.716",
"value": "International Standard Audiovisual Number"
},
{
"operation": "search",
"time": "44.915",
"value": "Contains... Lao Wang"
},
{
"operation": "Now on",
"time": "46.041",
"value": "undefined"
},
{
"operation": "search",
"time": "93.094",
"value": "Contains...isan society"
},
{
"operation": "Now on",
"time": "94.226",
"value": "undefined"
},
{
"operation": "hyperlink",
"time": "107.265",
"value": "Isan language"
},
{
"operation": "Now on",
"time": "112.639",
"value": "Isan language"
},
{
"operation": "hyperlink",
"time": "124.272",
"value": "Isan"
},
{
"operation": "Now on",
"time": "125.233",
"value": "Isan people"
},
{
"operation": "Page search",
"time": "138.35",
"value": "lao wang"
},
{
"operation": "hyperlink",
"time": "257.245",
"value": "Thai Cultural Mandates"
},
{
"operation": "Now on",
"time": "258.051",
"value": "Thai cultural mandates"
},
{
"operation": "Highlighting",
"time": "277.058",
"value": "Thai cultural mandates_item_14_37"
},
{
"operation": "Highlighting",
"time": "277.059",
"value": "Thai cultural mandates_title"
},
{
"operation": "Highlighting",
"time": "277.059",
"value": "Thai cultural mandates_section_13"
},
{
"operation": "Highlighting deleted",
"time": "279.876",
"value": "Thai cultural mandates_item_14_37"
},
{
"operation": "Highlighting deleted",
"time": "279.878",
"value": "Thai cultural mandates_title"
},
{
"operation": "Highlighting deleted",
"time": "279.881",
"value": "Thai cultural mandates_section_13"
},
{
"operation": "back-button-clicked",
"time": "292.74",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "292.871",
"value": "Isan people"
},
{
"operation": "hyperlink",
"time": "299.184",
"value": "Holy Men's Rebellions"
},
{
"operation": "Now on",
"time": "299.715",
"value": "Holy Man's Rebellion"
},
{
"operation": "back-button-clicked",
"time": "316.432",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "316.567",
"value": "Isan people"
},
{
"operation": "Highlighting",
"time": "380.67",
"value": "Isan people_sentence_38"
},
{
"operation": "finish",
"time": "410.834",
"value": "finish"
}
]
}
|
33851
|
The story tries to describe and reveal conflicts between the proprietors and peasants and provided further insight into some issues among the over 40's in South Korea.
|
{
"evidence": [
"The Taebaek Mountains_sentence_2",
"The Taebaek Mountains_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 story tries to describe and reveal conflicts between the proprietors and peasants and provided further insight into some issues among the over 40's in South Korea.\n",
"role": "user"
}
] |
Search terms not in claim
|
{
"ground_truth": {
"evidence": [
"The Taebaek Mountains_sentence_2",
"The Taebaek Mountains_sentence_4"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 5002,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.929",
"value": "?search="
},
{
"operation": "search",
"time": "21.527",
"value": "story tries to describe and reveal conflicts between the proprietors and peasants"
},
{
"operation": "Now on",
"time": "23.838",
"value": "undefined"
},
{
"operation": "hyperlink",
"time": "72.647",
"value": "Plot The Taebaek Mountains_section_0"
},
{
"operation": "Now on",
"time": "73.475",
"value": "The Taebaek Mountains#Plot The Taebaek Mountains section 0"
},
{
"operation": "Highlighting",
"time": "79.453",
"value": "The Taebaek Mountains_sentence_2"
},
{
"operation": "Highlighting",
"time": "91.601",
"value": "The Taebaek Mountains_sentence_4"
},
{
"operation": "finish",
"time": "102.915",
"value": "finish"
}
]
}
|
92754
|
Wayne Embry, a two-time all-Mid-American Conference center team captain, studied in Miami University in Oxford, Ohio wherein he tried to join Alpha Phi Alpha fraternity—the first intercollegiate historically African American Greek-lettered fraternity which was founded in 1906—but he did not push through with it.
|
{
"evidence": [
"Wayne Embry_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:\nWayne Embry, a two-time all-Mid-American Conference center team captain, studied in Miami University in Oxford, Ohio wherein he tried to join Alpha Phi Alpha fraternity—the first intercollegiate historically African American Greek-lettered fraternity which was founded in 1906—but he did not push through with it.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Wayne Embry_sentence_5"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 5003,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.546",
"value": "?search="
},
{
"operation": "search",
"time": "15.551",
"value": "Wayne Embry"
},
{
"operation": "Now on",
"time": "16.526",
"value": "Wayne Embry"
},
{
"operation": "Page search",
"time": "33.05",
"value": "all-mid"
},
{
"operation": "Highlighting",
"time": "37.461",
"value": "Wayne Embry_sentence_6"
},
{
"operation": "Page search",
"time": "50.554",
"value": "captain"
},
{
"operation": "Highlighting",
"time": "69.035",
"value": "Wayne Embry_sentence_5"
},
{
"operation": "Highlighting deleted",
"time": "83.052",
"value": "Wayne Embry_sentence_6"
},
{
"operation": "finish",
"time": "92.185",
"value": "finish"
}
]
}
|
exfever_6006
|
The Hateful Eight is a 2015 film starring an Italian actor, often playing supporting villainous characters of unstable nature.
|
{
"evidence": [
"The Hateful Eight",
"Bruce Dern"
],
"label": "REFUTE"
}
|
exfever
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nThe Hateful Eight is a 2015 film starring an Italian actor, often playing supporting villainous characters of unstable nature.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"The Hateful Eight",
"Bruce Dern"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 5004,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
73316
|
Michigan State Normal Normalites played against Olivet College in October 17 and won 28-3.
|
{
"evidence": [
"Eastern Michigan Eagles football_sentence_134"
],
"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:\nMichigan State Normal Normalites played against Olivet College in October 17 and won 28-3.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Eastern Michigan Eagles football_sentence_134"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 5005,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.652",
"value": "?search="
},
{
"operation": "search",
"time": "16.247",
"value": "Michigan State Normal Normalites football"
},
{
"operation": "Now on",
"time": "21.281",
"value": "Eastern Michigan Eagles football"
},
{
"operation": "search",
"time": "45.725",
"value": "Michigan State Normal Normalites 2017"
},
{
"operation": "Now on",
"time": "46.874",
"value": "undefined"
},
{
"operation": "hyperlink",
"time": "55.244",
"value": "Eastern Michigan Eagles football"
},
{
"operation": "Now on",
"time": "60.765",
"value": "Eastern Michigan Eagles football"
},
{
"operation": "search",
"time": "176.008",
"value": "Michigan State Normal Normalites vs Olivet College"
},
{
"operation": "Now on",
"time": "176.709",
"value": "undefined"
},
{
"operation": "hyperlink",
"time": "182.412",
"value": "Eastern Michigan Eagles football"
},
{
"operation": "Now on",
"time": "183.015",
"value": "Eastern Michigan Eagles football"
},
{
"operation": "Page search",
"time": "188.771",
"value": "Olivet"
},
{
"operation": "Highlighting",
"time": "198.063",
"value": "Eastern Michigan Eagles football_sentence_134"
},
{
"operation": "Highlighting deleted",
"time": "199.996",
"value": "Eastern Michigan Eagles football_sentence_134"
},
{
"operation": "search",
"time": "321.403",
"value": "Normal Normalities"
},
{
"operation": "Now on",
"time": "322.638",
"value": "undefined"
},
{
"operation": "search",
"time": "340.366",
"value": "2017 Normal Normalities"
},
{
"operation": "Now on",
"time": "342.007",
"value": "undefined"
},
{
"operation": "search",
"time": "359.007",
"value": "Michigan State Normal Normalites football"
},
{
"operation": "Now on",
"time": "360.131",
"value": "Eastern Michigan Eagles football"
},
{
"operation": "search",
"time": "400.261",
"value": "Michigan State Normal Normalites men's basketball"
},
{
"operation": "Now on",
"time": "413.139",
"value": "Eastern Michigan Eagles men's basketball"
},
{
"operation": "Page search",
"time": "424.289",
"value": "Olivet"
},
{
"operation": "search",
"time": "443.529",
"value": "Michigan State Normal Normalites football 2017"
},
{
"operation": "Now on",
"time": "444.036",
"value": "undefined"
},
{
"operation": "back-button-clicked",
"time": "449.215",
"value": "back-button-clicked"
},
{
"operation": "back-button-clicked",
"time": "453.34",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "453.631",
"value": "Eastern Michigan Eagles football"
},
{
"operation": "search",
"time": "465.358",
"value": "Michigan State Normal Normalites football, 1926"
},
{
"operation": "Now on",
"time": "466.417",
"value": "1926 Michigan State Normal Normalites football team"
},
{
"operation": "search",
"time": "498.657",
"value": "Michigan State Normal Normalites football October 17"
},
{
"operation": "Now on",
"time": "500.126",
"value": "undefined"
},
{
"operation": "hyperlink",
"time": "506.868",
"value": "1911 Michigan State Normal Normalites football team"
},
{
"operation": "Now on",
"time": "507.413",
"value": "1911 Michigan State Normal Normalites football team"
},
{
"operation": "search",
"time": "516.486",
"value": "Michigan State Normal Normalites football, 1922"
},
{
"operation": "Now on",
"time": "517.326",
"value": "1922 Michigan State Normal Normalites football team"
},
{
"operation": "search",
"time": "523.285",
"value": "Michigan State Normal Normalites football, 1921"
},
{
"operation": "Now on",
"time": "524.108",
"value": "1921 Michigan State Normal Normalites football team"
},
{
"operation": "search",
"time": "529.652",
"value": "Michigan State Normal Normalites football, 1923"
},
{
"operation": "Now on",
"time": "530.552",
"value": "1923 Michigan State Normal Normalites football team"
},
{
"operation": "search",
"time": "545.647",
"value": "Michigan State Normal Normalites football, 1904"
},
{
"operation": "Now on",
"time": "546.765",
"value": "Michigan State Normal Normalites football, 1904"
},
{
"operation": "search",
"time": "577.526",
"value": "Michigan State Normal Normalites football, 1923"
},
{
"operation": "Now on",
"time": "578.167",
"value": "Michigan State Normal Normalites football, 1923"
},
{
"operation": "search",
"time": "594.11",
"value": "Michigan State Normal Normalites football, 1926"
},
{
"operation": "Now on",
"time": "594.9",
"value": "Michigan State Normal Normalites football, 1926"
},
{
"operation": "hyperlink",
"time": "604.542",
"value": "Michigan State Normal Normalites football, 1926"
},
{
"operation": "Now on",
"time": "605.404",
"value": "undefined"
},
{
"operation": "search",
"time": "609.463",
"value": "Michigan State Normal Normalites football"
},
{
"operation": "Now on",
"time": "610.227",
"value": "Michigan State Normal Normalites football"
},
{
"operation": "hyperlink",
"time": "657.077",
"value": "Liberty"
},
{
"operation": "Now on",
"time": "661.986",
"value": "Liberty Flames football"
},
{
"operation": "search",
"time": "678.045",
"value": "Michigan State Normal Normalites 28-3"
},
{
"operation": "Now on",
"time": "679.951",
"value": "undefined"
},
{
"operation": "hyperlink",
"time": "697.372",
"value": "1904 Michigan State Normal Normalites football team"
},
{
"operation": "Now on",
"time": "698.036",
"value": "1904 Michigan State Normal Normalites football team"
},
{
"operation": "search",
"time": "743.095",
"value": "Michigan State Normal Normalites football"
},
{
"operation": "Now on",
"time": "744.932",
"value": "Eastern Michigan Eagles football"
},
{
"operation": "Page search",
"time": "755.994",
"value": "Olivet"
},
{
"operation": "search",
"time": "783.477",
"value": "Michigan State Normal Normalites 1906"
},
{
"operation": "Now on",
"time": "784.353",
"value": "undefined"
},
{
"operation": "hyperlink",
"time": "786.084",
"value": "1906 Michigan State Normal Normalites football team"
},
{
"operation": "Now on",
"time": "786.809",
"value": "1906 Michigan State Normal Normalites football team"
},
{
"operation": "search",
"time": "818.727",
"value": "Michigan State Normal Normalites football"
},
{
"operation": "Now on",
"time": "819.673",
"value": "Michigan State Normal Normalites football"
},
{
"operation": "Page search",
"time": "822.16",
"value": "Olivet"
},
{
"operation": "Highlighting",
"time": "839.637",
"value": "Eastern Michigan Eagles football_sentence_133"
},
{
"operation": "Highlighting deleted",
"time": "841.62",
"value": "Eastern Michigan Eagles football_sentence_133"
},
{
"operation": "search",
"time": "867.127",
"value": "Michigan Intercollegiate Athletic Association conference champinoships"
},
{
"operation": "Now on",
"time": "869.746",
"value": "undefined"
},
{
"operation": "search",
"time": "878.891",
"value": "Michigan Intercollegiate Athletic Association"
},
{
"operation": "Now on",
"time": "882.798",
"value": "Michigan Intercollegiate Athletic Association"
},
{
"operation": "search",
"time": "906.158",
"value": "Michigan State Normal Normalites football"
},
{
"operation": "Now on",
"time": "907.083",
"value": "Michigan State Normal Normalites football"
},
{
"operation": "Page search",
"time": "912.259",
"value": "olivet"
},
{
"operation": "Highlighting",
"time": "920.99",
"value": "Eastern Michigan Eagles football_sentence_134"
},
{
"operation": "search",
"time": "983.728",
"value": "Eastern Michigan Yearly results"
},
{
"operation": "Now on",
"time": "985.896",
"value": "undefined"
},
{
"operation": "search",
"time": "1041.139",
"value": "eastern michigan football conference"
},
{
"operation": "Now on",
"time": "1042.637",
"value": "undefined"
},
{
"operation": "hyperlink",
"time": "1047.164",
"value": "Mid-American Conference"
},
{
"operation": "Now on",
"time": "1049.913",
"value": "Mid-American Conference"
},
{
"operation": "search",
"time": "1204.566",
"value": "Olivet"
},
{
"operation": "Now on",
"time": "1205.383",
"value": "Olivet"
},
{
"operation": "search",
"time": "1222.687",
"value": "Michigan State Normal Normalites football"
},
{
"operation": "Now on",
"time": "1223.605",
"value": "Michigan State Normal Normalites football"
},
{
"operation": "Page search",
"time": "1250.551",
"value": "olivet"
},
{
"operation": "search",
"time": "1310.308",
"value": "Mid-American Conference Championship"
},
{
"operation": "Now on",
"time": "1311.093",
"value": "Mid-American Conference Championship"
},
{
"operation": "hyperlink",
"time": "1316.587",
"value": "MAC Championship Game"
},
{
"operation": "Now on",
"time": "1318.39",
"value": "MAC Football Championship Game"
},
{
"operation": "search",
"time": "1341.313",
"value": "Normal Normalites vs Olivet"
},
{
"operation": "Now on",
"time": "1342.672",
"value": "undefined"
},
{
"operation": "search",
"time": "1350.273",
"value": "Olivet football"
},
{
"operation": "Now on",
"time": "1351.216",
"value": "undefined"
},
{
"operation": "hyperlink",
"time": "1355.628",
"value": "1910 Olivet football team"
},
{
"operation": "Now on",
"time": "1356.171",
"value": "1910 Olivet football team"
},
{
"operation": "back-button-clicked",
"time": "1361.621",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "1362.363",
"value": "undefined"
},
{
"operation": "hyperlink",
"time": "1364.454",
"value": "1909 Olivet football team"
},
{
"operation": "Now on",
"time": "1364.971",
"value": "1909 Olivet football team"
},
{
"operation": "back-button-clicked",
"time": "1367.069",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "1367.859",
"value": "undefined"
},
{
"operation": "hyperlink",
"time": "1369.685",
"value": "1907 Olivet football team"
},
{
"operation": "Now on",
"time": "1370.357",
"value": "1907 Olivet football team"
},
{
"operation": "back-button-clicked",
"time": "1375.924",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "1376.654",
"value": "undefined"
},
{
"operation": "hyperlink",
"time": "1378.661",
"value": "1894 Olivet football team"
},
{
"operation": "Now on",
"time": "1379.304",
"value": "1894 Olivet football team"
},
{
"operation": "back-button-clicked",
"time": "1393.173",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "1393.935",
"value": "undefined"
},
{
"operation": "finish",
"time": "1424.372",
"value": "finish"
}
]
}
|
exfever_1697
|
Murder on the Orient Express was the first film directed by Kenneth Branagh.
|
{
"evidence": [
"Murder on the Orient Express (2017 film)",
"Johnny Depp",
"Academy Awards"
],
"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:\nMurder on the Orient Express was the first film directed by Kenneth Branagh.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Murder on the Orient Express (2017 film)",
"Johnny Depp",
"Academy Awards"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 5006,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_2403
|
The Host is a 2013 American romantic science fiction thriller film directed and adapted from Stephenie Meyer's 2008 novel by a screenwriter, producer, and director from an island country in the southwestern Pacific Ocean.
|
{
"evidence": [
"The Host (2013 film)",
"Andrew Niccol",
"New Zealand"
],
"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 Host is a 2013 American romantic science fiction thriller film directed and adapted from Stephenie Meyer's 2008 novel by a screenwriter, producer, and director from an island country in the southwestern Pacific Ocean.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"The Host (2013 film)",
"Andrew Niccol",
"New Zealand"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 5007,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_377
|
Penny Dreadful's executive producers include a Barbadian stage director.
|
{
"evidence": [
"Penny Dreadful (TV series)",
"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:\nPenny Dreadful's executive producers include a Barbadian stage director.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Penny Dreadful (TV series)",
"Sam Mendes"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 5008,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
22760
|
The Kirtlebridge rail crash took place on October 2, 1872 at Caledonian Railway, Kirtlebridge, Dumfries and Galloway, Southern Scotland, when the shunting took place without the signalman's authority.
|
{
"evidence": [
"Kirtlebridge rail crash_sentence_0",
"Kirtlebridge rail crash_sentence_5",
"Kirtlebridge rail crash_sentence_2",
"Kirtlebridge rail crash_sentence_3",
"Kirtlebridge railway station_sentence_0"
],
"label": "NOT ENOUGH INFO"
}
|
feverous
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nThe Kirtlebridge rail crash took place on October 2, 1872 at Caledonian Railway, Kirtlebridge, Dumfries and Galloway, Southern Scotland, when the shunting took place without the signalman's authority.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Kirtlebridge rail crash_sentence_0",
"Kirtlebridge rail crash_sentence_5",
"Kirtlebridge rail crash_sentence_2",
"Kirtlebridge rail crash_sentence_3",
"Kirtlebridge railway station_sentence_0"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 5009,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.562",
"value": "?search="
},
{
"operation": "search",
"time": "14.505",
"value": "Yuba City bus disaster"
},
{
"operation": "Now on",
"time": "15.401",
"value": "Yuba City bus disaster"
},
{
"operation": "back-button-clicked",
"time": "115.684",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "115.714",
"value": "?search="
},
{
"operation": "search",
"time": "123.391",
"value": "Kirtlebridge rail crash"
},
{
"operation": "Now on",
"time": "124.172",
"value": "Kirtlebridge rail crash"
},
{
"operation": "Highlighting",
"time": "149.27",
"value": "Kirtlebridge rail crash_sentence_0"
},
{
"operation": "Highlighting",
"time": "152.791",
"value": "Kirtlebridge rail crash_sentence_5"
},
{
"operation": "Highlighting",
"time": "174.752",
"value": "Kirtlebridge rail crash_sentence_2"
},
{
"operation": "Highlighting",
"time": "179.307",
"value": "Kirtlebridge rail crash_sentence_3"
},
{
"operation": "Page search",
"time": "187.268",
"value": "signal"
},
{
"operation": "Page search",
"time": "217.211",
"value": "curtlebridge"
},
{
"operation": "hyperlink",
"time": "247.66",
"value": "Caledonian Railway"
},
{
"operation": "Now on",
"time": "248.592",
"value": "Caledonian Railway"
},
{
"operation": "search",
"time": "256.26",
"value": "dumfries"
},
{
"operation": "Now on",
"time": "257.488",
"value": "Dumfries"
},
{
"operation": "back-button-clicked",
"time": "259.711",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "259.77",
"value": "Caledonian Railway"
},
{
"operation": "Page search",
"time": "263.453",
"value": "dumfries"
},
{
"operation": "back-button-clicked",
"time": "294.106",
"value": "back-button-clicked"
},
{
"operation": "back-button-clicked",
"time": "294.109",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "294.145",
"value": "Kirtlebridge rail crash"
},
{
"operation": "hyperlink",
"time": "302.598",
"value": "Kirtlebridge railway station"
},
{
"operation": "Now on",
"time": "303.393",
"value": "Kirtlebridge railway station"
},
{
"operation": "Highlighting",
"time": "316.697",
"value": "Kirtlebridge railway station_sentence_1"
},
{
"operation": "Page search",
"time": "321.187",
"value": "dumfries"
},
{
"operation": "Highlighting",
"time": "327.831",
"value": "Kirtlebridge railway station_sentence_0"
},
{
"operation": "Highlighting deleted",
"time": "329.794",
"value": "Kirtlebridge railway station_sentence_1"
},
{
"operation": "back-button-clicked",
"time": "334.958",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "334.988",
"value": "Kirtlebridge rail crash"
},
{
"operation": "finish",
"time": "352.839",
"value": "finish"
}
]
}
|
exfever_1753
|
Funny Girl was the first biographical musical comedy-drama film directed by William Wyler.
|
{
"evidence": [
"Funny Girl (film)",
"Comedian"
],
"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:\nFunny Girl was the first biographical musical comedy-drama film directed by William Wyler.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Funny Girl (film)",
"Comedian"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 5010,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_8950
|
"Hot Right Now" is a single featuring singer Rita Ora by a British producer best known for making electronic music, as a second single from his third studio album, which was released on October 1, 2012.
|
{
"evidence": [
"Hot Right Now",
"Nextlevelism",
"DJ Fresh"
],
"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:\n\"Hot Right Now\" is a single featuring singer Rita Ora by a British producer best known for making electronic music, as a second single from his third studio album, which was released on October 1, 2012.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Hot Right Now",
"Nextlevelism",
"DJ Fresh"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 5011,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_4261
|
The Fosters is the first American family drama television series.
|
{
"evidence": [
"The Fosters (2013 TV series)",
"Peter Paige",
"Say Uncle (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:\nThe Fosters is the first American family drama television series.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"The Fosters (2013 TV series)",
"Peter Paige",
"Say Uncle (film)"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 5012,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_5713
|
Fantastic Beasts and Where to Find Them is a 2016 fantasy series directed by an English film director, producer, and writer who has directed feature films, short films, and television productions.
|
{
"evidence": [
"Fantastic Beasts and Where to Find Them (film)",
"David Yates"
],
"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:\nFantastic Beasts and Where to Find Them is a 2016 fantasy series directed by an English film director, producer, and writer who has directed feature films, short films, and television productions.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Fantastic Beasts and Where to Find Them (film)",
"David Yates"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 5013,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
66061
|
"Basket Case" by Green Day, an American rock band formed in 1987 in East Bay, California, was nominated for nine awards in 1995.
|
{
"evidence": [
"Basket Case (song)_sentence_0",
"Basket Case (song)_sentence_28",
"Green Day_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:\n\"Basket Case\" by Green Day, an American rock band formed in 1987 in East Bay, California, was nominated for nine awards in 1995.\n",
"role": "user"
}
] |
Multi-hop Reasoning
|
{
"ground_truth": {
"evidence": [
"Basket Case (song)_sentence_0",
"Basket Case (song)_sentence_28",
"Green Day_sentence_3"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 5014,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.737",
"value": "?search="
},
{
"operation": "search",
"time": "42.592",
"value": "Basket Case"
},
{
"operation": "Now on",
"time": "43.778",
"value": "Basket Case"
},
{
"operation": "hyperlink",
"time": "49.294",
"value": "\"Basket Case\" (song)"
},
{
"operation": "Now on",
"time": "50.471",
"value": "Basket Case (song)"
},
{
"operation": "Highlighting",
"time": "52.967",
"value": "Basket Case (song)_sentence_0"
},
{
"operation": "Highlighting",
"time": "67.935",
"value": "Basket Case (song)_sentence_19"
},
{
"operation": "Highlighting",
"time": "67.935",
"value": "Basket Case (song)_section_2"
},
{
"operation": "Highlighting",
"time": "84.239",
"value": "Basket Case (song)_sentence_28"
},
{
"operation": "hyperlink",
"time": "98.583",
"value": "Green Day"
},
{
"operation": "Now on",
"time": "99.601",
"value": "Green Day"
},
{
"operation": "Highlighting",
"time": "107.923",
"value": "Green Day_sentence_3"
},
{
"operation": "finish",
"time": "273.67",
"value": "finish"
}
]
}
|
63162
|
North Korean athlete Om Yun-chol currently owns three records in the 56kg men's weightlifting events, which have been contested three times as optional sport, at the Summer Universiade
|
{
"evidence": [
"Om Yun-chol_sentence_1",
"Om Yun-chol_sentence_4",
"Om Yun-chol_sentence_6"
],
"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:\nNorth Korean athlete Om Yun-chol currently owns three records in the 56kg men's weightlifting events, which have been contested three times as optional sport, at the Summer Universiade\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Om Yun-chol_sentence_1",
"Om Yun-chol_sentence_4",
"Om Yun-chol_sentence_6"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 5015,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.986",
"value": "?search="
},
{
"operation": "search",
"time": "14.167",
"value": "Om Yun-chol"
},
{
"operation": "Now on",
"time": "15.201",
"value": "Om Yun-chol"
},
{
"operation": "Highlighting",
"time": "20.009",
"value": "Om Yun-chol_sentence_1"
},
{
"operation": "Highlighting",
"time": "100.734",
"value": "Om Yun-chol_sentence_4"
},
{
"operation": "hyperlink",
"time": "132.757",
"value": "men's 56 kg event"
},
{
"operation": "Now on",
"time": "133.747",
"value": "Weightlifting at the 2012 Summer Olympics – Men's 56 kg"
},
{
"operation": "back-button-clicked",
"time": "146.67",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "146.789",
"value": "Om Yun-chol"
},
{
"operation": "search",
"time": "194.638",
"value": "Summer Universiade"
},
{
"operation": "Now on",
"time": "197.312",
"value": "Universiade"
},
{
"operation": "back-button-clicked",
"time": "210.148",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "210.264",
"value": "Om Yun-chol"
},
{
"operation": "Highlighting",
"time": "230.489",
"value": "Om Yun-chol_sentence_6"
},
{
"operation": "finish",
"time": "236.271",
"value": "finish"
}
]
}
|
29266
|
PAX2, or the so-called paired box gene 2, directs the transgene expression of the mid-hindbrain and developing kidney and has five distinct MHB-specific enhancers in its upstream region.
|
{
"evidence": [
"PAX2_sentence_17"
],
"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:\nPAX2, or the so-called paired box gene 2, directs the transgene expression of the mid-hindbrain and developing kidney and has five distinct MHB-specific enhancers in its upstream region.\n",
"role": "user"
}
] |
Numerical Reasoning
|
{
"ground_truth": {
"evidence": [
"PAX2_sentence_17"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 5016,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.285",
"value": "?search="
},
{
"operation": "search",
"time": "6.276",
"value": "PAX2"
},
{
"operation": "Now on",
"time": "7.713",
"value": "PAX2"
},
{
"operation": "Highlighting",
"time": "12.944",
"value": "PAX2_sentence_0"
},
{
"operation": "Page search",
"time": "24.096",
"value": "transgene"
},
{
"operation": "Highlighting",
"time": "44.269",
"value": "PAX2_sentence_16"
},
{
"operation": "Highlighting",
"time": "56.815",
"value": "PAX2_sentence_17"
},
{
"operation": "finish",
"time": "72.006",
"value": "finish"
}
]
}
|
exfever_5232
|
The existence of God, with a wide variety of arguments for and against, can be categorized as the branch of biology that studies the fundamental nature of reality.
|
{
"evidence": [
"Existence of God",
"Metaphysics"
],
"label": "REFUTE"
}
|
exfever
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nThe existence of God, with a wide variety of arguments for and against, can be categorized as the branch of biology that studies the fundamental nature of reality.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Existence of God",
"Metaphysics"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 5017,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
5230
|
John Wheeley Gough Gutch was born at Bristol in 1809, the son of John Mathew Gutch (British journalist and historian), and was educated as a surgeon at the infirmary there and he became a member of the Royal College of Surgeons.
|
{
"evidence": [
"John Wheeley Gough Gutch_sentence_2",
"John Wheeley Gough Gutch_sentence_3",
"John Mathew Gutch_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:\nJohn Wheeley Gough Gutch was born at Bristol in 1809, the son of John Mathew Gutch (British journalist and historian), and was educated as a surgeon at the infirmary there and he became a member of the Royal College of Surgeons.\n",
"role": "user"
}
] |
Multi-hop Reasoning
|
{
"ground_truth": {
"evidence": [
"John Wheeley Gough Gutch_sentence_2",
"John Wheeley Gough Gutch_sentence_3",
"John Mathew Gutch_sentence_0"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 5018,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.256",
"value": "?search="
},
{
"operation": "search",
"time": "17.476",
"value": "John Wheeley Gough Gutch"
},
{
"operation": "Now on",
"time": "18.781",
"value": "John Wheeley Gough Gutch"
},
{
"operation": "Highlighting",
"time": "221.511",
"value": "John Wheeley Gough Gutch_sentence_2"
},
{
"operation": "Highlighting",
"time": "225.089",
"value": "John Wheeley Gough Gutch_sentence_3"
},
{
"operation": "hyperlink",
"time": "231.444",
"value": "John Mathew Gutch"
},
{
"operation": "Now on",
"time": "232.502",
"value": "John Mathew Gutch"
},
{
"operation": "Highlighting",
"time": "238.33",
"value": "John Mathew Gutch_sentence_0"
},
{
"operation": "back-button-clicked",
"time": "264.21",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "264.403",
"value": "John Wheeley Gough Gutch"
},
{
"operation": "finish",
"time": "332.01",
"value": "finish"
}
]
}
|
60169
|
Gavin Francis books are all under the fiction genre.
|
{
"evidence": [
"Gavin Francis_sentence_0",
"Gavin Francis_sentence_8",
"Gavin Francis_sentence_9"
],
"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:\nGavin Francis books are all under the fiction genre.\n",
"role": "user"
}
] |
Multi-hop Reasoning
|
{
"ground_truth": {
"evidence": [
"Gavin Francis_sentence_0",
"Gavin Francis_sentence_8",
"Gavin Francis_sentence_9"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 5019,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.19",
"value": "?search="
},
{
"operation": "search",
"time": "8.465",
"value": "Gavin Francis"
},
{
"operation": "Now on",
"time": "9.328",
"value": "Gavin Francis"
},
{
"operation": "Highlighting",
"time": "11.903",
"value": "Gavin Francis_sentence_0"
},
{
"operation": "Highlighting",
"time": "23.433",
"value": "Gavin Francis_sentence_8"
},
{
"operation": "Highlighting",
"time": "27.052",
"value": "Gavin Francis_sentence_9"
},
{
"operation": "finish",
"time": "40.832",
"value": "finish"
}
]
}
|
exfever_6464
|
The Falling is a 2014 British mystery drama film starring Florence Pugh and an English actress who led a character in the HBO epic medieval fantasy drama series Game of Thrones (2011–2019).
|
{
"evidence": [
"The Falling (2014 film)",
"Maisie Williams"
],
"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 Falling is a 2014 British mystery drama film starring Florence Pugh and an English actress who led a character in the HBO epic medieval fantasy drama series Game of Thrones (2011–2019).\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"The Falling (2014 film)",
"Maisie Williams"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 5020,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
57098
|
The Queue automaton is very different from a Turing machine since a queue machine can be simulated more easily by a multi-tape Turing machine, a variant of the Turing machine that utilizes several tapes.
|
{
"evidence": [
"Queue automaton_sentence_1",
"Queue automaton_sentence_3",
"Queue automaton_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:\nThe Queue automaton is very different from a Turing machine since a queue machine can be simulated more easily by a multi-tape Turing machine, a variant of the Turing machine that utilizes several tapes.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Queue automaton_sentence_1",
"Queue automaton_sentence_3",
"Queue automaton_sentence_5"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 5021,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "2.034",
"value": "?search="
},
{
"operation": "search",
"time": "12.31",
"value": "Queue automaton"
},
{
"operation": "Now on",
"time": "13.565",
"value": "Queue automaton"
},
{
"operation": "Highlighting",
"time": "16.952",
"value": "Queue automaton_sentence_1"
},
{
"operation": "Highlighting",
"time": "40.917",
"value": "Queue automaton_sentence_3"
},
{
"operation": "Highlighting",
"time": "47.917",
"value": "Queue automaton_sentence_4"
},
{
"operation": "Highlighting",
"time": "59.707",
"value": "Queue automaton_sentence_5"
},
{
"operation": "Highlighting deleted",
"time": "60.495",
"value": "Queue automaton_sentence_4"
},
{
"operation": "finish",
"time": "110.77",
"value": "finish"
}
]
}
|
62249
|
Their highest point is the Kamenec (Steinberg, 1,072 m n.m.) in the Pohorská foothills and in the municipality of Pohorská Ves and after die Wende, a nature park (a nature park, or sometimes natural park) was established called the Novohradské hory with many natural attractions like the Terčino údolí, Žofínský prales (Sophien-Urwald) and Hojná Voda (Heilbrunn).
|
{
"evidence": [
"Gratzen Mountains_sentence_5",
"Gratzen Mountains_sentence_8",
"Nature park_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:\nTheir highest point is the Kamenec (Steinberg, 1,072 m n.m.) in the Pohorská foothills and in the municipality of Pohorská Ves and after die Wende, a nature park (a nature park, or sometimes natural park) was established called the Novohradské hory with many natural attractions like the Terčino údolí, Žofínský prales (Sophien-Urwald) and Hojná Voda (Heilbrunn).\n",
"role": "user"
}
] |
Entity Disambiguation
|
{
"ground_truth": {
"evidence": [
"Gratzen Mountains_sentence_5",
"Gratzen Mountains_sentence_8",
"Nature park_sentence_0"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 5022,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.152",
"value": "?search="
},
{
"operation": "search",
"time": "34.819",
"value": "Contains...Their highest point is the Kamenec"
},
{
"operation": "Now on",
"time": "35.629",
"value": "undefined"
},
{
"operation": "hyperlink",
"time": "39.395",
"value": "Gratzen Mountains"
},
{
"operation": "Now on",
"time": "40.158",
"value": "Gratzen Mountains"
},
{
"operation": "Highlighting",
"time": "52.979",
"value": "Gratzen Mountains_sentence_5"
},
{
"operation": "Highlighting",
"time": "78.499",
"value": "Gratzen Mountains_sentence_8"
},
{
"operation": "hyperlink",
"time": "79.485",
"value": "nature park"
},
{
"operation": "Now on",
"time": "80.378",
"value": "Nature park"
},
{
"operation": "Highlighting",
"time": "83.135",
"value": "Nature park_sentence_0"
},
{
"operation": "finish",
"time": "107.726",
"value": "finish"
}
]
}
|
17614
|
Dr John William Ballantyne was a German who made advances in antenatal pathology and midwifery while serving as a Fellow of the Royal Society of Edinburgh.
|
{
"evidence": [
"John William Ballantyne_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:\nDr John William Ballantyne was a German who made advances in antenatal pathology and midwifery while serving as a Fellow of the Royal Society of Edinburgh.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"John William Ballantyne_sentence_0"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 5023,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.598",
"value": "?search="
},
{
"operation": "search",
"time": "6.849",
"value": "John William Ballantyne"
},
{
"operation": "Now on",
"time": "7.374",
"value": "John William Ballantyne"
},
{
"operation": "Highlighting",
"time": "9.88",
"value": "John William Ballantyne_sentence_0"
},
{
"operation": "finish",
"time": "82.567",
"value": "finish"
}
]
}
|
exfever_2825
|
Lymelife is the first comedy-drama film directed by Derick Martini.
|
{
"evidence": [
"Lymelife",
"Long Island"
],
"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:\nLymelife is the first comedy-drama film directed by Derick Martini.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Lymelife",
"Long Island"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 5024,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
88656
|
Kraft Heinz, an American company formed when Kraft Foods and Heinz merged, acquired Unilever in 2018.
|
{
"evidence": [
"Kraft Heinz_sentence_11",
"Kraft Heinz_sentence_12",
"Kraft Heinz_sentence_13"
],
"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:\nKraft Heinz, an American company formed when Kraft Foods and Heinz merged, acquired Unilever in 2018.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Kraft Heinz_sentence_11",
"Kraft Heinz_sentence_12",
"Kraft Heinz_sentence_13"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 5025,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.145",
"value": "?search="
},
{
"operation": "search",
"time": "13.666",
"value": "Kraft Heinz"
},
{
"operation": "Now on",
"time": "14.922",
"value": "Kraft Heinz"
},
{
"operation": "Highlighting",
"time": "19.391",
"value": "Kraft Heinz_sentence_0"
},
{
"operation": "Page search",
"time": "95.585",
"value": "unilever"
},
{
"operation": "Highlighting",
"time": "121.575",
"value": "Kraft Heinz_sentence_11"
},
{
"operation": "Highlighting",
"time": "128.899",
"value": "Kraft Heinz_sentence_12"
},
{
"operation": "hyperlink",
"time": "205.51",
"value": "Unilever"
},
{
"operation": "Now on",
"time": "207.464",
"value": "Unilever"
},
{
"operation": "Page search",
"time": "225.758",
"value": "kraft heinz"
},
{
"operation": "Page search",
"time": "226.498",
"value": "kraft heinz"
},
{
"operation": "back-button-clicked",
"time": "308.496",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "308.576",
"value": "Kraft Heinz"
},
{
"operation": "Highlighting",
"time": "360.222",
"value": "Kraft Heinz_sentence_13"
},
{
"operation": "Highlighting deleted",
"time": "365.994",
"value": "Kraft Heinz_sentence_0"
},
{
"operation": "finish",
"time": "373.923",
"value": "finish"
}
]
}
|
exfever_4490
|
Colin Edward Quinn's notable film work includes his role as Dooey in a 1999 American comedy film.
|
{
"evidence": [
"Colin Quinn",
"A Night at the Roxbury"
],
"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:\nColin Edward Quinn's notable film work includes his role as Dooey in a 1999 American comedy film.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Colin Quinn",
"A Night at the Roxbury"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 5026,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_6911
|
Jessica Lange is a British actress who won?an award given to an actress who has delivered an outstanding performance in a leading role in a film released that year.
|
{
"evidence": [
"Jessica Lange",
"Academy Award for Best Actress"
],
"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:\nJessica Lange is a British actress who won?an award given to an actress who has delivered an outstanding performance in a leading role in a film released that year.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Jessica Lange",
"Academy Award for Best Actress"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 5027,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
37258
|
Murray Aynsley Hill, a suburb of Christchurch, New Zealand was named after early settler Hugh Murray-Aynsley but he never owned the land.
|
{
"evidence": [
"Murray Aynsley Hill_sentence_0",
"Murray Aynsley Hill_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:\nMurray Aynsley Hill, a suburb of Christchurch, New Zealand was named after early settler Hugh Murray-Aynsley but he never owned the land.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Murray Aynsley Hill_sentence_0",
"Murray Aynsley Hill_sentence_2"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 5028,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.748",
"value": "?search="
},
{
"operation": "search",
"time": "19.237",
"value": "Murray Aynsley Hill"
},
{
"operation": "Now on",
"time": "19.895",
"value": "Murray Aynsley Hill"
},
{
"operation": "Highlighting",
"time": "25.118",
"value": "Murray Aynsley Hill_sentence_0"
},
{
"operation": "Highlighting",
"time": "30.738",
"value": "Murray Aynsley Hill_sentence_2"
},
{
"operation": "finish",
"time": "90.524",
"value": "finish"
}
]
}
|
exfever_459
|
Ruth Negga is an American actress who has received several major nominations for various awards, including the Oscars, given annually by a professional honorary organization with the stated goal of advancing the arts and sciences of motion picture.
|
{
"evidence": [
"Ruth Negga",
"Academy Awards",
"Academy of Motion Picture Arts and Sciences"
],
"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:\nRuth Negga is an American actress who has received several major nominations for various awards, including the Oscars, given annually by a professional honorary organization with the stated goal of advancing the arts and sciences of motion picture.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Ruth Negga",
"Academy Awards",
"Academy of Motion Picture Arts and Sciences"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 5029,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
22633
|
ODECO, founded by Alden J. "Doc" Laborde and John Hayward in 1957, was eventually acquired by Diamond Offshore Drilling (a company headquartered in Houston, Texas).
|
{
"evidence": [
"ODECO_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:\nODECO, founded by Alden J. \"Doc\" Laborde and John Hayward in 1957, was eventually acquired by Diamond Offshore Drilling (a company headquartered in Houston, Texas).\n",
"role": "user"
}
] |
Numerical Reasoning
|
{
"ground_truth": {
"evidence": [
"ODECO_sentence_2"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 5030,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.698",
"value": "?search="
},
{
"operation": "search",
"time": "11.03",
"value": "ODECO"
},
{
"operation": "Now on",
"time": "11.761",
"value": "ODECO"
},
{
"operation": "Highlighting",
"time": "23.99",
"value": "ODECO_sentence_2"
},
{
"operation": "finish",
"time": "33.982",
"value": "finish"
}
]
}
|
exfever_5119
|
Sabrina Annlynn Carpenter is an American singer whose breakthrough starring in the film Girl Meets World by an American pay television channel that serves as the flagship property of Disney Branded Television.
|
{
"evidence": [
"Sabrina Carpenter",
"Disney Channel"
],
"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:\nSabrina Annlynn Carpenter is an American singer whose breakthrough starring in the film Girl Meets World by an American pay television channel that serves as the flagship property of Disney Branded Television.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Sabrina Carpenter",
"Disney Channel"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 5031,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_2098
|
How to Train Your Dragon is a 2010 American computer-animated fantasy film with a film genre in which the protagonist is thrust into a series of events that typically involve violence and physical feats.
|
{
"evidence": [
"How to Train Your Dragon (film)",
"Action 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:\nHow to Train Your Dragon is a 2010 American computer-animated fantasy film with a film genre in which the protagonist is thrust into a series of events that typically involve violence and physical feats.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"How to Train Your Dragon (film)",
"Action film"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 5032,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
6833
|
In August 2011, Jack announced on social media that he was close to finishing a five-track EP, but later released another statement that he recently finished a four-track EP of new songs.
|
{
"evidence": [
"Just Jack_sentence_19",
"Just Jack_sentence_20"
],
"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 August 2011, Jack announced on social media that he was close to finishing a five-track EP, but later released another statement that he recently finished a four-track EP of new songs.\n",
"role": "user"
}
] |
Search terms not in claim
|
{
"ground_truth": {
"evidence": [
"Just Jack_sentence_19",
"Just Jack_sentence_20"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 5033,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.06",
"value": "?search="
},
{
"operation": "search",
"time": "173.586",
"value": "just jack"
},
{
"operation": "Now on",
"time": "174.882",
"value": "Just jack"
},
{
"operation": "Highlighting",
"time": "184.006",
"value": "Just Jack_sentence_19"
},
{
"operation": "Highlighting",
"time": "188.926",
"value": "Just Jack_sentence_20"
},
{
"operation": "finish",
"time": "195.574",
"value": "finish"
}
]
}
|
exfever_7073
|
Dame Julia Mary Walters is an English actress, author, and comedian, who rose to prominence by playing the title role in a 1983 British film whose genre of dramatic work combines elements of comedy and action.
|
{
"evidence": [
"Julie Walters",
"Educating Rita (film)",
"Comedy-drama"
],
"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:\nDame Julia Mary Walters is an English actress, author, and comedian, who rose to prominence by playing the title role in a 1983 British film whose genre of dramatic work combines elements of comedy and action.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Julie Walters",
"Educating Rita (film)",
"Comedy-drama"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 5034,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
53479
|
Vernonia noveboracensis is a herbaceous plant, or vascular plants that have no persistent woody stems above ground, with alternate, simple leaves, on stiff, greenish purple stems and has purple flowers borne in summer and fall.
|
{
"evidence": [
"Vernonia noveboracensis_sentence_2",
"Vernonia noveboracensis_sentence_3",
"Herbaceous plant_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:\nVernonia noveboracensis is a herbaceous plant, or vascular plants that have no persistent woody stems above ground, with alternate, simple leaves, on stiff, greenish purple stems and has purple flowers borne in summer and fall.\n",
"role": "user"
}
] |
Multi-hop Reasoning
|
{
"ground_truth": {
"evidence": [
"Vernonia noveboracensis_sentence_2",
"Vernonia noveboracensis_sentence_3",
"Herbaceous plant_sentence_2"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 5035,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.332",
"value": "?search="
},
{
"operation": "search",
"time": "17.023",
"value": "Vernonia noveboracensis"
},
{
"operation": "Now on",
"time": "17.951",
"value": "Vernonia noveboracensis"
},
{
"operation": "Highlighting",
"time": "52.949",
"value": "Vernonia noveboracensis_sentence_2"
},
{
"operation": "Highlighting",
"time": "79.092",
"value": "Vernonia noveboracensis_sentence_3"
},
{
"operation": "hyperlink",
"time": "102.205",
"value": "herbaceous plant"
},
{
"operation": "Now on",
"time": "103.216",
"value": "Herbaceous plant"
},
{
"operation": "Highlighting",
"time": "110.848",
"value": "Herbaceous plant_sentence_2"
},
{
"operation": "finish",
"time": "152.876",
"value": "finish"
}
]
}
|
85519
|
Manufactured by Stadler Rail, Variobahn, a German-designed model of articulated low-floor tram and light rail vehicle, comprises 25.71% of the total number of trams operating in Tramz.
|
{
"evidence": [
"Stadler Variobahn_sentence_0",
"Stadler Variobahn_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:\nManufactured by Stadler Rail, Variobahn, a German-designed model of articulated low-floor tram and light rail vehicle, comprises 25.71% of the total number of trams operating in Tramz.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Stadler Variobahn_sentence_0",
"Stadler Variobahn_sentence_1"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 5036,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.082",
"value": "?search="
},
{
"operation": "search",
"time": "13.254",
"value": "Variobahn for TfL"
},
{
"operation": "Now on",
"time": "14.237",
"value": "Stadler Variobahn"
},
{
"operation": "Highlighting",
"time": "21.222",
"value": "Stadler Variobahn_sentence_0"
},
{
"operation": "Highlighting",
"time": "37.043",
"value": "Stadler Variobahn_sentence_1"
},
{
"operation": "Page search",
"time": "49.559",
"value": "tramz"
},
{
"operation": "Page search",
"time": "66.391",
"value": "25.71"
},
{
"operation": "Page search",
"time": "138.271",
"value": "25"
},
{
"operation": "finish",
"time": "151.457",
"value": "finish"
}
]
}
|
exfever_3991
|
Keith Urban is a New Zealand-born Australian musician, singer, guitarist, and songwriter who performs a song nominated to honor the best in film and television of 2012.
|
{
"evidence": [
"Keith Urban",
"For You (Keith Urban song)",
"70th Golden Globe Awards"
],
"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:\nKeith Urban is a New Zealand-born Australian musician, singer, guitarist, and songwriter who performs a song nominated to honor the best in film and television of 2012.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Keith Urban",
"For You (Keith Urban song)",
"70th Golden Globe Awards"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 5037,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
3229
|
Each National Olympic Committee (NOC) can enter a maximum of 2 competitors per gender and each of the five continents must be represented in each gender but if no athletes qualify the China will also qualify.
|
{
"evidence": [
"Archery at the 2020 Summer Olympics_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:\nEach National Olympic Committee (NOC) can enter a maximum of 2 competitors per gender and each of the five continents must be represented in each gender but if no athletes qualify the China will also qualify.\n",
"role": "user"
}
] |
Search terms not in claim
|
{
"ground_truth": {
"evidence": [
"Archery at the 2020 Summer Olympics_sentence_5"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 5038,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "search",
"time": "15.056",
"value": "National Olympic Committee"
},
{
"operation": "Page search",
"time": "22.734",
"value": "compe"
},
{
"operation": "Page search",
"time": "23.878",
"value": "compe"
},
{
"operation": "Page search",
"time": "49.819",
"value": "maximum"
},
{
"operation": "search",
"time": "95.746",
"value": "Archery at the 2020 Summer Olympics"
},
{
"operation": "Page search",
"time": "100.269",
"value": "maximum"
},
{
"operation": "Highlighting",
"time": "105.966",
"value": "Archery at the 2020 Summer Olympics_sentence_5"
},
{
"operation": "Page search",
"time": "125.088",
"value": "china"
},
{
"operation": "Page search",
"time": "164.483",
"value": "china"
},
{
"operation": "Page search",
"time": "164.874",
"value": "china"
},
{
"operation": "Highlighting",
"time": "181.274",
"value": "Archery at the 2020 Summer Olympics_sentence_5"
},
{
"operation": "Page search",
"time": "218.504",
"value": "gender"
},
{
"operation": "finish",
"time": "243.335",
"value": "finish"
}
]
}
|
exfever_4288
|
John Cusack is an American actor who has starred in numerous films, including Better Off Dead. Better Off Dead is a 1985 American comedy film written and directed by Will Smith.
|
{
"evidence": [
"John Cusack",
"Better Off Dead (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:\nJohn Cusack is an American actor who has starred in numerous films, including Better Off Dead. Better Off Dead is a 1985 American comedy film written and directed by Will Smith.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"John Cusack",
"Better Off Dead (film)"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 5039,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
29741
|
A63 autoroute is a motorway in southwest France that is part of the Autoroute des Estuaires and connects the border with Spain at Biriatou.
|
{
"evidence": [
"A63 autoroute_sentence_0",
"A63 autoroute_sentence_4",
"A63 autoroute_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:\nA63 autoroute is a motorway in southwest France that is part of the Autoroute des Estuaires and connects the border with Spain at Biriatou.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"A63 autoroute_sentence_0",
"A63 autoroute_sentence_4",
"A63 autoroute_sentence_5"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 5040,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.171",
"value": "?search="
},
{
"operation": "search",
"time": "49.611",
"value": "A63 autoroute"
},
{
"operation": "Now on",
"time": "50.89",
"value": "A63 autoroute"
},
{
"operation": "Highlighting",
"time": "54.013",
"value": "A63 autoroute_sentence_0"
},
{
"operation": "Highlighting",
"time": "73.079",
"value": "A63 autoroute_sentence_4"
},
{
"operation": "Highlighting",
"time": "77.182",
"value": "A63 autoroute_sentence_5"
},
{
"operation": "finish",
"time": "82.951",
"value": "finish"
}
]
}
|
45017
|
Rebecca Cox Jackson married Samuel S. Jackson (not to be confused with actor Samuel L. Jackson) but got divorced after he failed to teach her how to read and write.
|
{
"evidence": [
"Rebecca Cox Jackson_sentence_2",
"Rebecca Cox Jackson_sentence_4",
"Samuel L. Jackson_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:\nRebecca Cox Jackson married Samuel S. Jackson (not to be confused with actor Samuel L. Jackson) but got divorced after he failed to teach her how to read and write.\n",
"role": "user"
}
] |
Multi-hop Reasoning
|
{
"ground_truth": {
"evidence": [
"Rebecca Cox Jackson_sentence_2",
"Rebecca Cox Jackson_sentence_4",
"Samuel L. Jackson_sentence_2"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 5041,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.801",
"value": "?search="
},
{
"operation": "search",
"time": "3.658",
"value": "Rebecca Cox Jackson"
},
{
"operation": "Now on",
"time": "4.513",
"value": "Rebecca Cox Jackson"
},
{
"operation": "Highlighting",
"time": "8.501",
"value": "Rebecca Cox Jackson_sentence_2"
},
{
"operation": "Highlighting",
"time": "15.297",
"value": "Rebecca Cox Jackson_sentence_4"
},
{
"operation": "search",
"time": "24.908",
"value": "Samuel L. Jackson"
},
{
"operation": "Now on",
"time": "26.362",
"value": "Samuel L. Jackson"
},
{
"operation": "Highlighting",
"time": "28.435",
"value": "Samuel L. Jackson_sentence_2"
},
{
"operation": "finish",
"time": "51.839",
"value": "finish"
}
]
}
|
85288
|
Released in Japan in 2011, New Album was the fifteenth studio album by the band Boris, which was formed in 1992 in Tokyo, Japan.
|
{
"evidence": [
"Boris (band)_sentence_0",
"New Album_sentence_2",
"New Album_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:\nReleased in Japan in 2011, New Album was the fifteenth studio album by the band Boris, which was formed in 1992 in Tokyo, Japan.\n",
"role": "user"
}
] |
Multi-hop Reasoning
|
{
"ground_truth": {
"evidence": [
"Boris (band)_sentence_0",
"New Album_sentence_2",
"New Album_sentence_3"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 5042,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.385",
"value": "?search="
},
{
"operation": "search",
"time": "13.228",
"value": "Boris"
},
{
"operation": "Now on",
"time": "13.969",
"value": "Boris"
},
{
"operation": "hyperlink",
"time": "19.178",
"value": "Boris (band)"
},
{
"operation": "Now on",
"time": "20.178",
"value": "Boris (band)"
},
{
"operation": "Highlighting",
"time": "24.601",
"value": "Boris (band)_sentence_0"
},
{
"operation": "hyperlink",
"time": "39.563",
"value": "Boris discography"
},
{
"operation": "Now on",
"time": "40.72",
"value": "Boris discography"
},
{
"operation": "hyperlink",
"time": "57.686",
"value": "New Album"
},
{
"operation": "Now on",
"time": "58.278",
"value": "New Album"
},
{
"operation": "Highlighting",
"time": "63.84",
"value": "New Album_sentence_2"
},
{
"operation": "Highlighting",
"time": "72.233",
"value": "New Album_sentence_3"
},
{
"operation": "finish",
"time": "84.407",
"value": "finish"
}
]
}
|
63808
|
The Judicial District of Lima, the Judicial District with the most operations in the whole nation, was created by the decree of August 4, 1824 and was established December 22, 1824 under the presidency of Simón Bolívar, a Venezuelan military and political leader who led what are currently the countries of Venezuela, Bolivia, Colombia, Ecuador, Peru, and Panama to independence from the Spanish Empire.
|
{
"evidence": [
"Judicial District of Lima_sentence_5",
"Judicial District of Lima_sentence_4",
"Simón Bolívar_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:\nThe Judicial District of Lima, the Judicial District with the most operations in the whole nation, was created by the decree of August 4, 1824 and was established December 22, 1824 under the presidency of Simón Bolívar, a Venezuelan military and political leader who led what are currently the countries of Venezuela, Bolivia, Colombia, Ecuador, Peru, and Panama to independence from the Spanish Empire.\n",
"role": "user"
}
] |
Multi-hop Reasoning
|
{
"ground_truth": {
"evidence": [
"Judicial District of Lima_sentence_5",
"Judicial District of Lima_sentence_4",
"Simón Bolívar_sentence_3"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 5043,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.936",
"value": "?search="
},
{
"operation": "search",
"time": "16.261",
"value": "Judicial District of Lima"
},
{
"operation": "Now on",
"time": "18.014",
"value": "Judicial District of Lima"
},
{
"operation": "Highlighting",
"time": "24.5",
"value": "Judicial District of Lima_sentence_5"
},
{
"operation": "Highlighting",
"time": "32.028",
"value": "Judicial District of Lima_sentence_4"
},
{
"operation": "Highlighting",
"time": "63.484",
"value": "Judicial District of Lima_sentence_0"
},
{
"operation": "Highlighting deleted",
"time": "64.656",
"value": "Judicial District of Lima_sentence_0"
},
{
"operation": "Highlighting",
"time": "76.162",
"value": "Judicial District of Lima_sentence_2"
},
{
"operation": "Highlighting deleted",
"time": "76.954",
"value": "Judicial District of Lima_sentence_2"
},
{
"operation": "hyperlink",
"time": "91.246",
"value": "Simón Bolívar"
},
{
"operation": "Now on",
"time": "97.464",
"value": "Simón Bolívar"
},
{
"operation": "Highlighting",
"time": "104.145",
"value": "Simón Bolívar_sentence_3"
},
{
"operation": "finish",
"time": "127.638",
"value": "finish"
}
]
}
|
51289
|
Javier Andorrà Julia is a forward who plays for the B-team of FC Andorra, he is Andorran and was born on 7 June 1985.
|
{
"evidence": [
"Xavi Andorrà_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:\nJavier Andorrà Julia is a forward who plays for the B-team of FC Andorra, he is Andorran and was born on 7 June 1985.\n",
"role": "user"
}
] |
Search terms not in claim
|
{
"ground_truth": {
"evidence": [
"Xavi Andorrà_sentence_0"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 5044,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.916",
"value": "?search="
},
{
"operation": "search",
"time": "207.886",
"value": "Xavi Andorrà"
},
{
"operation": "Now on",
"time": "208.699",
"value": "Xavi Andorrà"
},
{
"operation": "Highlighting",
"time": "240.854",
"value": "Xavi Andorrà_sentence_0"
},
{
"operation": "finish",
"time": "348.569",
"value": "finish"
}
]
}
|
62199
|
Early artists of the abstract illusionism movement inlucded Ronald Davis and Al Held, an American artist.
|
{
"evidence": [
"Abstract illusionism_sentence_7",
"Al Held_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:\nEarly artists of the abstract illusionism movement inlucded Ronald Davis and Al Held, an American artist.\n",
"role": "user"
}
] |
Multi-hop Reasoning
|
{
"ground_truth": {
"evidence": [
"Abstract illusionism_sentence_7",
"Al Held_sentence_0"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 5045,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.133",
"value": "?search="
},
{
"operation": "search",
"time": "64.383",
"value": "abstract illusionism"
},
{
"operation": "Now on",
"time": "65.387",
"value": "Abstract illusionism"
},
{
"operation": "Page search",
"time": "71.78",
"value": "artists"
},
{
"operation": "Page search",
"time": "82.345",
"value": "ronald davis"
},
{
"operation": "Highlighting",
"time": "86.738",
"value": "Abstract illusionism_sentence_7"
},
{
"operation": "hyperlink",
"time": "94.754",
"value": "Al Held"
},
{
"operation": "Now on",
"time": "95.468",
"value": "Al Held"
},
{
"operation": "Highlighting",
"time": "106.006",
"value": "Al Held_sentence_0"
},
{
"operation": "finish",
"time": "318.315",
"value": "finish"
}
]
}
|
31541
|
Iowa State Capitol, a 275 ft. high building, was the tallest building in Iowa from 1884 to 1924; then from 1991 up to the present, 801 Grand in Des Moines, which is 630 ft. high with 45 floors, holds the title.
|
{
"evidence": [
"Iowa State Capitol_sentence_57"
],
"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:\nIowa State Capitol, a 275 ft. high building, was the tallest building in Iowa from 1884 to 1924; then from 1991 up to the present, 801 Grand in Des Moines, which is 630 ft. high with 45 floors, holds the title.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Iowa State Capitol_sentence_57"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 5046,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.063",
"value": "undefined"
},
{
"operation": "Now on",
"time": "24.613",
"value": "undefined"
},
{
"operation": "search",
"time": "125.226",
"value": "iowa state capitol"
},
{
"operation": "Now on",
"time": "126.384",
"value": "Iowa State Capitol"
},
{
"operation": "Highlighting",
"time": "258.848",
"value": "Iowa State Capitol_sentence_57"
},
{
"operation": "search",
"time": "376.425",
"value": "801 grand"
},
{
"operation": "Now on",
"time": "377.571",
"value": "801 Grand"
},
{
"operation": "finish",
"time": "481.613",
"value": "finish"
}
]
}
|
32322
|
Bulimba A power station was designed in 1928 by the Council to expand the power which had an initial capacity of 18.75 MW and expanded to 93.75 MW upon completion.
|
{
"evidence": [
"History of electricity supply in Brisbane_sentence_21"
],
"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:\nBulimba A power station was designed in 1928 by the Council to expand the power which had an initial capacity of 18.75 MW and expanded to 93.75 MW upon completion.\n",
"role": "user"
}
] |
Search terms not in claim
|
{
"ground_truth": {
"evidence": [
"History of electricity supply in Brisbane_sentence_21"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 5047,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.674",
"value": "?search="
},
{
"operation": "search",
"time": "13.888",
"value": "Contains...Bulimba A power station"
},
{
"operation": "Now on",
"time": "15.55",
"value": "undefined"
},
{
"operation": "hyperlink",
"time": "20.921",
"value": "History of electricity supply in Brisbane"
},
{
"operation": "Now on",
"time": "22.188",
"value": "History of electricity supply in Brisbane"
},
{
"operation": "Page search",
"time": "26.999",
"value": "Bulimba A power station"
},
{
"operation": "Highlighting",
"time": "31.695",
"value": "History of electricity supply in Brisbane_sentence_21"
},
{
"operation": "search",
"time": "84.159",
"value": "Contains...Bulimba A power station"
},
{
"operation": "Now on",
"time": "85.515",
"value": "undefined"
},
{
"operation": "search",
"time": "108.155",
"value": "\"Bulimba A power station\""
},
{
"operation": "Now on",
"time": "108.863",
"value": "undefined"
},
{
"operation": "hyperlink",
"time": "111.344",
"value": "History of electricity supply in Brisbane"
},
{
"operation": "Now on",
"time": "112.407",
"value": "History of electricity supply in Brisbane"
},
{
"operation": "Page search",
"time": "124.892",
"value": "Bulimba A"
},
{
"operation": "finish",
"time": "156.427",
"value": "finish"
}
]
}
|
709
|
In the list of Unification of Norway (860–872), Vestfold (current electoral district in Eastern Norway) gained four victories while Viking gained three victories.
|
{
"evidence": [
"Vestfold_sentence_0",
"Vestfold_sentence_25"
],
"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 list of Unification of Norway (860–872), Vestfold (current electoral district in Eastern Norway) gained four victories while Viking gained three victories.\n",
"role": "user"
}
] |
Multi-hop Reasoning
|
{
"ground_truth": {
"evidence": [
"Vestfold_sentence_0",
"Vestfold_sentence_25"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 5048,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "search",
"time": "10.856",
"value": "Vestfold"
},
{
"operation": "Highlighting",
"time": "18.277",
"value": "Vestfold_sentence_0"
},
{
"operation": "hyperlink",
"time": "23.683",
"value": "2.1 Viking history Vestfold_section_2"
},
{
"operation": "Highlighting",
"time": "25.897",
"value": "Vestfold_sentence_25"
},
{
"operation": "search",
"time": "42.013",
"value": "unification of norway"
},
{
"operation": "finish",
"time": "55.991",
"value": "finish"
}
]
}
|
exfever_1115
|
Abraham Lincoln was an American lawyer and statesman who served as the 16th president of the United States.
|
{
"evidence": [
"Theodore Roosevelt",
"Assassination of William McKinley"
],
"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:\nAbraham Lincoln was an American lawyer and statesman who served as the 16th president of the United States.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Theodore Roosevelt",
"Assassination of William McKinley"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 5049,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
46265
|
Semra Bheropura is a village that offers a lot of jobs.
|
{
"evidence": [
"Semra Bheropura_sentence_0"
],
"label": "NOT ENOUGH INFO"
}
|
feverous
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nSemra Bheropura is a village that offers a lot of jobs.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Semra Bheropura_sentence_0"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 5050,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.02",
"value": "?search="
},
{
"operation": "search",
"time": "8.608",
"value": "Semra Bheropura"
},
{
"operation": "Now on",
"time": "9.523",
"value": "Semra Bheropura"
},
{
"operation": "Highlighting",
"time": "27.096",
"value": "Semra Bheropura_sentence_0"
},
{
"operation": "finish",
"time": "48.281",
"value": "finish"
}
]
}
|
exfever_8246
|
The Fate of the Furious is written by an American screenwriter with multiple screenplays in the Fox crime drama series that aired on Fox from May 22 to August 14, 2014.
|
{
"evidence": [
"The Fate of the Furious",
"Chris Morgan (writer)",
"Gang Related (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 Fate of the Furious is written by an American screenwriter with multiple screenplays in the Fox crime drama series that aired on Fox from May 22 to August 14, 2014.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"The Fate of the Furious",
"Chris Morgan (writer)",
"Gang Related (TV series)"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 5051,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_6014
|
Gabrielle Monique Union is known for her performance in a 2001 French romantic comedy drama film.
|
{
"evidence": [
"Gabrielle Union",
"The Brothers (2001 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:\nGabrielle Monique Union is known for her performance in a 2001 French romantic comedy drama film.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Gabrielle Union",
"The Brothers (2001 film)"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 5052,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_8580
|
Brock Edward Lesnar was born in Webster, South Dakota, on July 12, 1977.
|
{
"evidence": [
"Brock Lesnar",
"American football",
"Defense (sports)"
],
"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:\nBrock Edward Lesnar was born in Webster, South Dakota, on July 12, 1977.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Brock Lesnar",
"American football",
"Defense (sports)"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 5053,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_8395
|
David Lawrence Schwimmer is an American actor, comedian, director, and producer who starred in the 1998 American action film based on Stephen King's 1982 novella.
|
{
"evidence": [
"David Schwimmer",
"Apt Pupil (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:\nDavid Lawrence Schwimmer is an American actor, comedian, director, and producer who starred in the 1998 American action film based on Stephen King's 1982 novella.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"David Schwimmer",
"Apt Pupil (film)"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 5054,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
64040
|
Peter Tscherkassky was born in Vienna, Austria and attended the Primary School in Mistelbach.
|
{
"evidence": [
"Peter Tscherkassky_sentence_2",
"Peter Tscherkassky_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:\nPeter Tscherkassky was born in Vienna, Austria and attended the Primary School in Mistelbach.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Peter Tscherkassky_sentence_2",
"Peter Tscherkassky_sentence_3"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 5055,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "2.034",
"value": "?search="
},
{
"operation": "search",
"time": "9.615",
"value": "Peter Tscherkassky"
},
{
"operation": "Now on",
"time": "10.3",
"value": "Peter Tscherkassky"
},
{
"operation": "Highlighting",
"time": "87.165",
"value": "Peter Tscherkassky_sentence_2"
},
{
"operation": "Highlighting",
"time": "90.579",
"value": "Peter Tscherkassky_sentence_3"
},
{
"operation": "finish",
"time": "95.709",
"value": "finish"
}
]
}
|
exfever_1835
|
Peter Paul Cetera is an English musician best known for being the lead vocalist and bassist of an American rock band formed in Chicago, Illinois, that produced numerous top-40 hits over two decades.
|
{
"evidence": [
"Peter Cetera",
"Chicago (band)"
],
"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:\nPeter Paul Cetera is an English musician best known for being the lead vocalist and bassist of an American rock band formed in Chicago, Illinois, that produced numerous top-40 hits over two decades.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Peter Cetera",
"Chicago (band)"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 5056,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
75532
|
Michael Nyman Live is a 1994 album by Michael Nyman and the Michael Nyman Band whose early printings of the album cover listed the first three tracks erroneously as "Queen of the Night", "An Eye for Optical Theory", and "Chasing Sheep Is Best Left to Shepherds".
|
{
"evidence": [
"Michael Nyman Live_sentence_0",
"Michael Nyman Live_sentence_6"
],
"label": "SUPPORT"
}
|
feverous
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nMichael Nyman Live is a 1994 album by Michael Nyman and the Michael Nyman Band whose early printings of the album cover listed the first three tracks erroneously as \"Queen of the Night\", \"An Eye for Optical Theory\", and \"Chasing Sheep Is Best Left to Shepherds\".\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Michael Nyman Live_sentence_0",
"Michael Nyman Live_sentence_6"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 5057,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.032",
"value": "?search="
},
{
"operation": "search",
"time": "21.547",
"value": "Michael Nyman Live"
},
{
"operation": "Now on",
"time": "22.757",
"value": "Michael Nyman Live"
},
{
"operation": "Highlighting",
"time": "25.688",
"value": "Michael Nyman Live_sentence_0"
},
{
"operation": "Highlighting",
"time": "43.523",
"value": "Michael Nyman Live_sentence_6"
},
{
"operation": "finish",
"time": "66.848",
"value": "finish"
}
]
}
|
6521
|
Hamdi Salihi moved for the first time outside joining Panionios and played five games before moving back.
|
{
"evidence": [
"Hamdi Salihi_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:\nHamdi Salihi moved for the first time outside joining Panionios and played five games before moving back.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Hamdi Salihi_sentence_11"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 5058,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "9.148",
"value": "?search="
},
{
"operation": "search",
"time": "21.314",
"value": "hamdi salihi"
},
{
"operation": "Now on",
"time": "27.195",
"value": "Hamdi Salihi"
},
{
"operation": "Page search",
"time": "60.862",
"value": "panionios"
},
{
"operation": "Highlighting",
"time": "83.483",
"value": "Hamdi Salihi_sentence_10"
},
{
"operation": "Highlighting deleted",
"time": "89.133",
"value": "Hamdi Salihi_sentence_10"
},
{
"operation": "Highlighting",
"time": "94.761",
"value": "Hamdi Salihi_sentence_11"
},
{
"operation": "finish",
"time": "159.663",
"value": "finish"
}
]
}
|
25269
|
Suguru Egawa, who was nicknamed "The Monster", set a still-standing record of 60 strikeouts in National High School Baseball Invitational Tournament in the spring of 1973.
|
{
"evidence": [
"Suguru Egawa_sentence_4",
"Suguru Egawa_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:\nSuguru Egawa, who was nicknamed \"The Monster\", set a still-standing record of 60 strikeouts in National High School Baseball Invitational Tournament in the spring of 1973.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Suguru Egawa_sentence_4",
"Suguru Egawa_sentence_5"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 5059,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.151",
"value": "?search="
},
{
"operation": "search",
"time": "8.607",
"value": "Suguru Egawa"
},
{
"operation": "Now on",
"time": "9.557",
"value": "Suguru Egawa"
},
{
"operation": "Highlighting",
"time": "22.839",
"value": "Suguru Egawa_sentence_4"
},
{
"operation": "Highlighting",
"time": "24.271",
"value": "Suguru Egawa_sentence_5"
},
{
"operation": "finish",
"time": "46.023",
"value": "finish"
}
]
}
|
exfever_8281
|
Robert James Ritchie is a French singer who started his music career as a rapper and DJ, releasing his subsequent independent third studio album by Kid Rock and the first to feature his backing band Twisted Brown Trucker.
|
{
"evidence": [
"Kid Rock",
"Early Mornin' Stoned Pimp"
],
"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:\nRobert James Ritchie is a French singer who started his music career as a rapper and DJ, releasing his subsequent independent third studio album by Kid Rock and the first to feature his backing band Twisted Brown Trucker.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Kid Rock",
"Early Mornin' Stoned Pimp"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 5060,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_3728
|
Ryan John Seacrest is an Asian media personality and producer who hosted multiple media platforms, including an internationally syndicated, independent song countdown radio program created by Casey Kasem, Don Bustany, Tom Rounds, and Ron Jacobs.
|
{
"evidence": [
"Ryan Seacrest",
"American Top 40"
],
"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:\nRyan John Seacrest is an Asian media personality and producer who hosted multiple media platforms, including an internationally syndicated, independent song countdown radio program created by Casey Kasem, Don Bustany, Tom Rounds, and Ron Jacobs.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Ryan Seacrest",
"American Top 40"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 5061,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_605
|
Priyanka Chopra is one of the Indian actress who is a fan favorite in Hollywood.
|
{
"evidence": [
"Kaminey",
"Action 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:\nPriyanka Chopra is one of the Indian actress who is a fan favorite in Hollywood.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Kaminey",
"Action film"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 5062,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
8658
|
Humshaugh is one of the largest villages near Cilurnum, a Britannia province founded in 123 AD, which lies just off the military road.
|
{
"evidence": [
"Humshaugh_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:\nHumshaugh is one of the largest villages near Cilurnum, a Britannia province founded in 123 AD, which lies just off the military road.\n",
"role": "user"
}
] |
Numerical Reasoning
|
{
"ground_truth": {
"evidence": [
"Humshaugh_sentence_1"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 5063,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.674",
"value": "?search="
},
{
"operation": "search",
"time": "9.372",
"value": "Humshaugh"
},
{
"operation": "Now on",
"time": "10.579",
"value": "Humshaugh"
},
{
"operation": "Highlighting",
"time": "29.402",
"value": "Humshaugh_sentence_1"
},
{
"operation": "hyperlink",
"time": "97.42",
"value": "Cilurnum"
},
{
"operation": "Now on",
"time": "98.075",
"value": "Cilurnum"
},
{
"operation": "back-button-clicked",
"time": "104.542",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "104.8",
"value": "Humshaugh"
},
{
"operation": "hyperlink",
"time": "137.998",
"value": "Cilurnum"
},
{
"operation": "Now on",
"time": "138.877",
"value": "Cilurnum"
},
{
"operation": "back-button-clicked",
"time": "145.846",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "146.102",
"value": "Humshaugh"
},
{
"operation": "hyperlink",
"time": "149.035",
"value": "Hexham"
},
{
"operation": "Now on",
"time": "150.283",
"value": "Hexham"
},
{
"operation": "back-button-clicked",
"time": "157.334",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "157.648",
"value": "Humshaugh"
},
{
"operation": "finish",
"time": "163.692",
"value": "finish"
}
]
}
|
20991
|
Downs, born on June 9, 1959 and graduated from South Oak Cliff High School in Dallas, Texas, lettered in track and field, and was accepted a football scholarship from Rice University where he became a three-year starter at safety.
|
{
"evidence": [
"Michael Downs_sentence_2",
"Michael Downs_sentence_4",
"Michael Downs_sentence_6",
"Michael Downs_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:\nDowns, born on June 9, 1959 and graduated from South Oak Cliff High School in Dallas, Texas, lettered in track and field, and was accepted a football scholarship from Rice University where he became a three-year starter at safety.\n",
"role": "user"
}
] |
Entity Disambiguation
|
{
"ground_truth": {
"evidence": [
"Michael Downs_sentence_2",
"Michael Downs_sentence_4",
"Michael Downs_sentence_6",
"Michael Downs_sentence_0"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 5064,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.605",
"value": "?search="
},
{
"operation": "search",
"time": "14.161",
"value": "downs football"
},
{
"operation": "Now on",
"time": "15.181",
"value": "undefined"
},
{
"operation": "search",
"time": "30.751",
"value": "June 9, 1959 downs"
},
{
"operation": "Now on",
"time": "32.752",
"value": "undefined"
},
{
"operation": "search",
"time": "52.343",
"value": "June 9, 1959"
},
{
"operation": "Now on",
"time": "53.212",
"value": "undefined"
},
{
"operation": "hyperlink",
"time": "63.511",
"value": "Michael Downs"
},
{
"operation": "Now on",
"time": "64.288",
"value": "Michael Downs"
},
{
"operation": "Highlighting",
"time": "73.33",
"value": "Michael Downs_header_cell_0_1_0"
},
{
"operation": "Highlighting",
"time": "74.122",
"value": "Michael Downs_cell_0_1_1"
},
{
"operation": "Highlighting",
"time": "74.131",
"value": "Michael Downs_header_cell_0_1_0"
},
{
"operation": "Highlighting",
"time": "74.131",
"value": "Michael Downs_header_cell_0_0_0"
},
{
"operation": "Highlighting",
"time": "75.972",
"value": "Michael Downs_header_cell_0_3_0"
},
{
"operation": "Highlighting",
"time": "77.957",
"value": "Michael Downs_cell_0_3_1"
},
{
"operation": "Highlighting",
"time": "77.965",
"value": "Michael Downs_header_cell_0_3_0"
},
{
"operation": "Highlighting",
"time": "77.965",
"value": "Michael Downs_header_cell_0_2_0"
},
{
"operation": "Highlighting",
"time": "84.614",
"value": "Michael Downs_sentence_2"
},
{
"operation": "Highlighting",
"time": "88.883",
"value": "Michael Downs_sentence_4"
},
{
"operation": "Highlighting",
"time": "92.739",
"value": "Michael Downs_sentence_6"
},
{
"operation": "Highlighting deleted",
"time": "97.242",
"value": "Michael Downs_cell_0_1_1"
},
{
"operation": "Highlighting deleted",
"time": "97.248",
"value": "Michael Downs_header_cell_0_1_0"
},
{
"operation": "Highlighting deleted",
"time": "97.25",
"value": "Michael Downs_header_cell_0_0_0"
},
{
"operation": "Highlighting",
"time": "99.719",
"value": "Michael Downs_sentence_0"
},
{
"operation": "Highlighting deleted",
"time": "101.988",
"value": "Michael Downs_cell_0_3_1"
},
{
"operation": "Highlighting deleted",
"time": "101.995",
"value": "Michael Downs_header_cell_0_3_0"
},
{
"operation": "Highlighting deleted",
"time": "101.996",
"value": "Michael Downs_header_cell_0_2_0"
},
{
"operation": "finish",
"time": "253.541",
"value": "finish"
}
]
}
|
68664
|
Eagle Cave is an onyx cave that was discovered in 1849, opened to the public in 1938 and offers over 3,000 feet of passages for exploring across eight levels in Richland County, Wisconsin, United States.
|
{
"evidence": [
"Eagle Cave_sentence_0",
"Eagle Cave_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:\nEagle Cave is an onyx cave that was discovered in 1849, opened to the public in 1938 and offers over 3,000 feet of passages for exploring across eight levels in Richland County, Wisconsin, United States.\n",
"role": "user"
}
] |
Numerical Reasoning
|
{
"ground_truth": {
"evidence": [
"Eagle Cave_sentence_0",
"Eagle Cave_sentence_2"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 5065,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.148",
"value": "?search="
},
{
"operation": "search",
"time": "13.966",
"value": "Eagle Cave"
},
{
"operation": "Now on",
"time": "14.624",
"value": "Eagle Cave"
},
{
"operation": "Highlighting",
"time": "19.342",
"value": "Eagle Cave_sentence_0"
},
{
"operation": "Highlighting",
"time": "68.867",
"value": "Eagle Cave_sentence_2"
},
{
"operation": "hyperlink",
"time": "225.607",
"value": "Richland County, Wisconsin"
},
{
"operation": "Now on",
"time": "227.618",
"value": "Richland County, Wisconsin"
},
{
"operation": "search",
"time": "252.491",
"value": "3,000 feet of passages"
},
{
"operation": "Now on",
"time": "256.99",
"value": "undefined"
},
{
"operation": "search",
"time": "274.772",
"value": "eight levels eagle cave"
},
{
"operation": "Now on",
"time": "277.374",
"value": "undefined"
},
{
"operation": "search",
"time": "291.697",
"value": "eagle cave"
},
{
"operation": "Now on",
"time": "294.173",
"value": "Eagle Cave"
},
{
"operation": "finish",
"time": "365.686",
"value": "finish"
}
]
}
|
48377
|
India pale ale was not different in the past, in the mid-eighteenth century, it was mostly brewed with coke-fired malt which produced less smoke.
|
{
"evidence": [
"India pale ale_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:\nIndia pale ale was not different in the past, in the mid-eighteenth century, it was mostly brewed with coke-fired malt which produced less smoke.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"India pale ale_sentence_3"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 5066,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.67",
"value": "?search="
},
{
"operation": "search",
"time": "17.069",
"value": "India pale ale"
},
{
"operation": "Now on",
"time": "17.856",
"value": "India pale ale"
},
{
"operation": "Highlighting",
"time": "106.693",
"value": "India pale ale_sentence_3"
},
{
"operation": "finish",
"time": "119.314",
"value": "finish"
}
]
}
|
78683
|
Prévost reaction is a chemical reaction in which an alkene is converted by iodine and the silver salt of benzoic acid to a vicinal diol with anti sterochemistry and the reaction was discovered by the French chemist Charles Prevost.
|
{
"evidence": [
"Prévost reaction_sentence_0",
"Prévost reaction_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:\nPrévost reaction is a chemical reaction in which an alkene is converted by iodine and the silver salt of benzoic acid to a vicinal diol with anti sterochemistry and the reaction was discovered by the French chemist Charles Prevost.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Prévost reaction_sentence_0",
"Prévost reaction_sentence_1"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 5067,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.126",
"value": "?search="
},
{
"operation": "search",
"time": "6.529",
"value": "prevost reaction"
},
{
"operation": "Now on",
"time": "7.165",
"value": "Prévost reaction"
},
{
"operation": "Highlighting",
"time": "14.141",
"value": "Prévost reaction_sentence_0"
},
{
"operation": "Highlighting",
"time": "20.857",
"value": "Prévost reaction_sentence_1"
},
{
"operation": "finish",
"time": "28.923",
"value": "finish"
}
]
}
|
36562
|
Maria Fitzherbert (26 July 1756 – 27 March 1837)married Edward Weld who died two months later and was left destitute.
|
{
"evidence": [
"Maria Fitzherbert_sentence_12",
"Maria Fitzherbert_sentence_13"
],
"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:\nMaria Fitzherbert (26 July 1756 – 27 March 1837)married Edward Weld who died two months later and was left destitute.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Maria Fitzherbert_sentence_12",
"Maria Fitzherbert_sentence_13"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 5068,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.318",
"value": "?search="
},
{
"operation": "search",
"time": "12.517",
"value": "Maria Fitzherbert"
},
{
"operation": "Now on",
"time": "13.736",
"value": "Maria Fitzherbert"
},
{
"operation": "Highlighting",
"time": "25.908",
"value": "Maria Fitzherbert_sentence_3"
},
{
"operation": "Page search",
"time": "53.155",
"value": "Weld"
},
{
"operation": "Highlighting",
"time": "63.547",
"value": "Maria Fitzherbert_sentence_6"
},
{
"operation": "Highlighting",
"time": "66.349",
"value": "Maria Fitzherbert_sentence_7"
},
{
"operation": "page-search-reset",
"time": "96.54",
"value": "page-search-reset"
},
{
"operation": "Highlighting deleted",
"time": "102.724",
"value": "Maria Fitzherbert_sentence_7"
},
{
"operation": "Highlighting",
"time": "104.636",
"value": "Maria Fitzherbert_sentence_7"
},
{
"operation": "Highlighting deleted",
"time": "134.589",
"value": "Maria Fitzherbert_sentence_3"
},
{
"operation": "Highlighting deleted",
"time": "136.014",
"value": "Maria Fitzherbert_sentence_6"
},
{
"operation": "Highlighting deleted",
"time": "137.11",
"value": "Maria Fitzherbert_sentence_7"
},
{
"operation": "Highlighting",
"time": "153.246",
"value": "Maria Fitzherbert_sentence_12"
},
{
"operation": "Highlighting",
"time": "154.854",
"value": "Maria Fitzherbert_sentence_13"
},
{
"operation": "finish",
"time": "164.759",
"value": "finish"
}
]
}
|
exfever_941
|
Emily Blunt is a British actress who appeared in the 2007 British biographical-historical television film about the queen of the Iceni tribe.
|
{
"evidence": [
"Emily Blunt",
"Boudica (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:\nEmily Blunt is a British actress who appeared in the 2007 British biographical-historical television film about the queen of the Iceni tribe.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Emily Blunt",
"Boudica (film)"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 5069,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_4405
|
Klute is a 1978 American genre of film that originally flourished during the post-World War II era in the United States.
|
{
"evidence": [
"Klute",
"Neo-noir"
],
"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:\nKlute is a 1978 American genre of film that originally flourished during the post-World War II era in the United States.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Klute",
"Neo-noir"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 5070,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_3776
|
Splatoon 2 has a sequel that also releases on Nintendo Direct.
|
{
"evidence": [
"Splatoon 2",
"Nintendo Switch"
],
"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:\nSplatoon 2 has a sequel that also releases on Nintendo Direct.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Splatoon 2",
"Nintendo Switch"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 5071,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_1533
|
Artemis is the goddess of the hunt and the natural environments on Earth that are nonurbanized land not under extensive agricultural cultivation.
|
{
"evidence": [
"Artemis",
"Wilderness"
],
"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:\nArtemis is the goddess of the hunt and the natural environments on Earth that are nonurbanized land not under extensive agricultural cultivation.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Artemis",
"Wilderness"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 5072,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_3779
|
Paul von Hindenburg was a German Field Marshal and statesman who was appointed as Chancellor of Germany by an Austrian-born German politician who was the dictator of Germany from 1933 until his death in 1965.
|
{
"evidence": [
"Paul von Hindenburg",
"Adolf Hitler"
],
"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:\nPaul von Hindenburg was a German Field Marshal and statesman who was appointed as Chancellor of Germany by an Austrian-born German politician who was the dictator of Germany from 1933 until his death in 1965.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Paul von Hindenburg",
"Adolf Hitler"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 5073,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
37495
|
While Piłsudski dodged the bullets, the Lwów Province Governor, Kazimierz Grabowski, was wounded in the opening ceremony of the first Eastern Trade Fair; which by 1928, could boast neither 1,600 exhibitors nor 400 foreign firms.
|
{
"evidence": [
"Targi Wschodnie_sentence_6",
"Targi Wschodnie_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:\nWhile Piłsudski dodged the bullets, the Lwów Province Governor, Kazimierz Grabowski, was wounded in the opening ceremony of the first Eastern Trade Fair; which by 1928, could boast neither 1,600 exhibitors nor 400 foreign firms.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Targi Wschodnie_sentence_6",
"Targi Wschodnie_sentence_8"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 5074,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.578",
"value": "?search="
},
{
"operation": "search",
"time": "30.142",
"value": "Piłsudski dodged the bullets, the Lwów Province Governor, Kazimierz Grabowski, was wounded"
},
{
"operation": "Now on",
"time": "31.729",
"value": "undefined"
},
{
"operation": "hyperlink",
"time": "35.524",
"value": "Targi Wschodnie"
},
{
"operation": "Now on",
"time": "36.492",
"value": "Targi Wschodnie"
},
{
"operation": "Page search",
"time": "43.565",
"value": "bullet"
},
{
"operation": "Highlighting",
"time": "50.244",
"value": "Targi Wschodnie_sentence_7"
},
{
"operation": "Highlighting",
"time": "63.175",
"value": "Targi Wschodnie_sentence_6"
},
{
"operation": "Highlighting",
"time": "71.272",
"value": "Targi Wschodnie_sentence_8"
},
{
"operation": "Highlighting",
"time": "82.917",
"value": "Targi Wschodnie_sentence_8"
},
{
"operation": "Highlighting",
"time": "94.951",
"value": "Targi Wschodnie_sentence_6"
},
{
"operation": "Highlighting deleted",
"time": "139.476",
"value": "Targi Wschodnie_sentence_6"
},
{
"operation": "Highlighting deleted",
"time": "140.351",
"value": "Targi Wschodnie_sentence_8"
},
{
"operation": "Highlighting",
"time": "141.19",
"value": "Targi Wschodnie_sentence_6"
},
{
"operation": "Highlighting",
"time": "142.668",
"value": "Targi Wschodnie_sentence_8"
},
{
"operation": "finish",
"time": "145.853",
"value": "finish"
}
]
}
|
18669
|
Dragan Apic was loaded (sic) out to FMP for the rest of the 2017-2018 season; in early 2019 he signed a two-year deal with Lokomotiv Kuban; then in early 2020 he was loaned to San Pablo Burgos for the rest of the 2019-2020 season.
|
{
"evidence": [
"Dragan Apić_sentence_3",
"Dragan Apić_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:\nDragan Apic was loaded (sic) out to FMP for the rest of the 2017-2018 season; in early 2019 he signed a two-year deal with Lokomotiv Kuban; then in early 2020 he was loaned to San Pablo Burgos for the rest of the 2019-2020 season.\n",
"role": "user"
}
] |
Numerical Reasoning
|
{
"ground_truth": {
"evidence": [
"Dragan Apić_sentence_3",
"Dragan Apić_sentence_5"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 5075,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.999",
"value": "?search="
},
{
"operation": "search",
"time": "9.946",
"value": "Dragan Apic"
},
{
"operation": "Now on",
"time": "10.706",
"value": "Dragan Apić"
},
{
"operation": "Highlighting",
"time": "25.113",
"value": "Dragan Apić_sentence_3"
},
{
"operation": "Highlighting",
"time": "36.407",
"value": "Dragan Apić_sentence_5"
},
{
"operation": "finish",
"time": "45.355",
"value": "finish"
}
]
}
|
23446
|
Matsya is an avatar of Wishnu who appears as half horse/half man; he saves the world from cosmic deluge with the help of a boat made from knowledge, on which he rescues all living beings.
|
{
"evidence": [
"Matsya_sentence_4"
],
"label": "REFUTE"
}
|
feverous
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nMatsya is an avatar of Wishnu who appears as half horse/half man; he saves the world from cosmic deluge with the help of a boat made from knowledge, on which he rescues all living beings.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Matsya_sentence_4"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 5076,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.541",
"value": "?search="
},
{
"operation": "search",
"time": "6.617",
"value": "Matsya"
},
{
"operation": "Now on",
"time": "7.372",
"value": "Matsya"
},
{
"operation": "Highlighting",
"time": "12.122",
"value": "Matsya_sentence_3"
},
{
"operation": "Highlighting deleted",
"time": "18.444",
"value": "Matsya_sentence_3"
},
{
"operation": "Highlighting",
"time": "19.625",
"value": "Matsya_sentence_2"
},
{
"operation": "Highlighting deleted",
"time": "24.052",
"value": "Matsya_sentence_2"
},
{
"operation": "Highlighting",
"time": "26.071",
"value": "Matsya_sentence_1"
},
{
"operation": "Highlighting",
"time": "26.933",
"value": "Matsya_sentence_2"
},
{
"operation": "Highlighting deleted",
"time": "37.035",
"value": "Matsya_sentence_2"
},
{
"operation": "Highlighting deleted",
"time": "38.055",
"value": "Matsya_sentence_1"
},
{
"operation": "Highlighting",
"time": "41.987",
"value": "Matsya_sentence_4"
},
{
"operation": "Highlighting",
"time": "45.647",
"value": "Matsya_sentence_2"
},
{
"operation": "Highlighting deleted",
"time": "47.736",
"value": "Matsya_sentence_2"
},
{
"operation": "Highlighting deleted",
"time": "48.462",
"value": "Matsya_sentence_4"
},
{
"operation": "Highlighting",
"time": "49.799",
"value": "Matsya_sentence_1"
},
{
"operation": "Highlighting",
"time": "50.724",
"value": "Matsya_sentence_2"
},
{
"operation": "Highlighting",
"time": "51.407",
"value": "Matsya_sentence_4"
},
{
"operation": "Page search",
"time": "53.022",
"value": "horse"
},
{
"operation": "Highlighting deleted",
"time": "93.418",
"value": "Matsya_sentence_2"
},
{
"operation": "Highlighting deleted",
"time": "95.455",
"value": "Matsya_sentence_1"
},
{
"operation": "finish",
"time": "116.16",
"value": "finish"
}
]
}
|
exfever_2737
|
John Malkovich is an American actor who has appeared in more than 70 films, including a 1977 American epic coming-of-age war film directed by Steven Spielberg.
|
{
"evidence": [
"John Malkovich",
"Empire of the Sun (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:\nJohn Malkovich is an American actor who has appeared in more than 70 films, including a 1977 American epic coming-of-age war film directed by Steven Spielberg.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"John Malkovich",
"Empire of the Sun (film)"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 5077,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_3002
|
Bruce Willis is an American retired actor who began his career in the theatre, which is any professional theatre venue in New York City.
|
{
"evidence": [
"Bruce Willis",
"Off-Broadway"
],
"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:\nBruce Willis is an American retired actor who began his career in the theatre, which is any professional theatre venue in New York City.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Bruce Willis",
"Off-Broadway"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 5078,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_8269
|
Emma Stone was named as one of the 100 most influential people in the world by a news magazine based in the most populous city in the United States.
|
{
"evidence": [
"Emma Stone",
"Time (magazine)",
"New York City"
],
"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:\nEmma Stone was named as one of the 100 most influential people in the world by a news magazine based in the most populous city in the United States.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Emma Stone",
"Time (magazine)",
"New York City"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 5079,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_1128
|
Spider-Man 3 was Thomas Haden Church's first film as a villain.
|
{
"evidence": [
"Thomas Haden Church",
"Sandman (Marvel Comics)",
"Marvel Comics"
],
"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:\nSpider-Man 3 was Thomas Haden Church's first film as a villain.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Thomas Haden Church",
"Sandman (Marvel Comics)",
"Marvel Comics"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 5080,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_4220
|
A pet is a companion multicellular organism in the biological kingdom Animalia that is kept primarily for a person's company.
|
{
"evidence": [
"Pet",
"Animal"
],
"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 pet is a companion multicellular organism in the biological kingdom Animalia that is kept primarily for a person's company.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Pet",
"Animal"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 5081,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_1615
|
Selena Marie Gomez is a singer who rose to prominence for her lead role as Alex Russo in an American teen sitcom whose genre of television programming featured elements of the fantastic, including magic, that aired on Disney Channel.
|
{
"evidence": [
"Selena Gomez",
"Wizards of Waverly Place",
"Fantasy television"
],
"label": "SUPPORT"
}
|
exfever
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nSelena Marie Gomez is a singer who rose to prominence for her lead role as Alex Russo in an American teen sitcom whose genre of television programming featured elements of the fantastic, including magic, that aired on Disney Channel.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Selena Gomez",
"Wizards of Waverly Place",
"Fantasy television"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 5082,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_6708
|
PewDiePie is a Swedish novelist best known as an online contributor of information to any end-user/audience in specific contexts.
|
{
"evidence": [
"PewDiePie",
"Content creation"
],
"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:\nPewDiePie is a Swedish novelist best known as an online contributor of information to any end-user/audience in specific contexts.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"PewDiePie",
"Content creation"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 5083,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.