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_6140
|
Detroit Diesel Corporation is an American diesel engine manufacturer headquartered in a state in the Great Lakes region of the upper Midwestern United States.
|
{
"evidence": [
"Detroit Diesel",
"Michigan"
],
"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:\nDetroit Diesel Corporation is an American diesel engine manufacturer headquartered in a state in the Great Lakes region of the upper Midwestern United States.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Detroit Diesel",
"Michigan"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 484,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_62
|
Dredd earned just over $41 million at the box office on an estimated budget of $30–45 million.
|
{
"evidence": [
"Karl Urban",
"Dredd",
"Science fiction 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:\nDredd earned just over $41 million at the box office on an estimated budget of $30–45 million.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Karl Urban",
"Dredd",
"Science fiction film"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 485,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
38139
|
Restinga (has two ecoregions) is a forest with medium-sized trees found in eastern Brazil.
|
{
"evidence": [
"Restinga_sentence_1",
"Restinga_sentence_2",
"Restinga_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:\nRestinga (has two ecoregions) is a forest with medium-sized trees found in eastern Brazil.\n",
"role": "user"
}
] |
Multi-hop Reasoning
|
{
"ground_truth": {
"evidence": [
"Restinga_sentence_1",
"Restinga_sentence_2",
"Restinga_sentence_4"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 486,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.631",
"value": "?search="
},
{
"operation": "search",
"time": "5.078",
"value": "Restinga"
},
{
"operation": "Now on",
"time": "5.76",
"value": "Restinga"
},
{
"operation": "Page search",
"time": "7.721",
"value": "ecoregions"
},
{
"operation": "Highlighting",
"time": "10.23",
"value": "Restinga_sentence_4"
},
{
"operation": "Page search",
"time": "29.924",
"value": "medium-sized"
},
{
"operation": "Page search",
"time": "30.12",
"value": "medium-sized"
},
{
"operation": "Highlighting",
"time": "59.203",
"value": "Restinga_sentence_2"
},
{
"operation": "Page search",
"time": "67.892",
"value": "east"
},
{
"operation": "Highlighting",
"time": "77.425",
"value": "Restinga_sentence_1"
},
{
"operation": "Highlighting deleted",
"time": "80.497",
"value": "Restinga_sentence_2"
},
{
"operation": "Highlighting deleted",
"time": "81.315",
"value": "Restinga_sentence_4"
},
{
"operation": "Highlighting",
"time": "81.83",
"value": "Restinga_sentence_2"
},
{
"operation": "Highlighting",
"time": "82.38",
"value": "Restinga_sentence_4"
},
{
"operation": "finish",
"time": "116.183",
"value": "finish"
}
]
}
|
exfever_6212
|
Alkaline Trio from Chicago, Illinois is a Finnish band whose music is a broad genre of popular music that originated as "rock and roll" in the United States in the late 1940s.
|
{
"evidence": [
"Alkaline Trio",
"Rock music"
],
"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:\nAlkaline Trio from Chicago, Illinois is a Finnish band whose music is a broad genre of popular music that originated as \"rock and roll\" in the United States in the late 1940s.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Alkaline Trio",
"Rock music"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 487,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
44347
|
Prompt gamma neutron activation analysis measures using a gamma ray spectrometer, an instrument also used for analysis of elemental bodies in the Solar System.
|
{
"evidence": [
"Prompt gamma neutron activation analysis_sentence_5",
"Gamma-ray spectrometer_sentence_13"
],
"label": "SUPPORT"
}
|
feverous
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nPrompt gamma neutron activation analysis measures using a gamma ray spectrometer, an instrument also used for analysis of elemental bodies in the Solar System.\n",
"role": "user"
}
] |
Multi-hop Reasoning
|
{
"ground_truth": {
"evidence": [
"Prompt gamma neutron activation analysis_sentence_5",
"Gamma-ray spectrometer_sentence_13"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 488,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.071",
"value": "?search="
},
{
"operation": "search",
"time": "21.5",
"value": "Gamma-ray spectrometer"
},
{
"operation": "Now on",
"time": "22.683",
"value": "Gamma-ray spectrometer"
},
{
"operation": "search",
"time": "34.226",
"value": "Contains...gamma neutron activation analysis"
},
{
"operation": "Now on",
"time": "35.886",
"value": "undefined"
},
{
"operation": "hyperlink",
"time": "40",
"value": "Prompt gamma neutron activation analysis"
},
{
"operation": "Now on",
"time": "40.619",
"value": "Prompt gamma neutron activation analysis"
},
{
"operation": "Page search",
"time": "52.808",
"value": "spectrometer"
},
{
"operation": "Highlighting",
"time": "54.506",
"value": "Prompt gamma neutron activation analysis_sentence_5"
},
{
"operation": "hyperlink",
"time": "67.624",
"value": "gamma ray spectrometer"
},
{
"operation": "Now on",
"time": "69.37",
"value": "Gamma ray spectrometer"
},
{
"operation": "back-button-clicked",
"time": "71.492",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "71.718",
"value": "Prompt gamma neutron activation analysis"
},
{
"operation": "hyperlink",
"time": "86.65",
"value": "gamma ray spectrometer"
},
{
"operation": "Now on",
"time": "87.305",
"value": "Gamma-ray spectrometer"
},
{
"operation": "Page search",
"time": "96.758",
"value": "elemental"
},
{
"operation": "Highlighting",
"time": "110.525",
"value": "Gamma-ray spectrometer_sentence_13"
},
{
"operation": "finish",
"time": "130.797",
"value": "finish"
}
]
}
|
exfever_7694
|
The Aix-Marseille-Provence Metropolis is pronounced as métropole d'Aix-Marseille-Provence in a Romance language of the Indo-American family.
|
{
"evidence": [
"Aix-Marseille-Provence Metropolis",
"French language"
],
"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 Aix-Marseille-Provence Metropolis is pronounced as métropole d'Aix-Marseille-Provence in a Romance language of the Indo-American family.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Aix-Marseille-Provence Metropolis",
"French language"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 489,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_2103
|
Jimi Hendrix is the greatest guitarist of all time.
|
{
"evidence": [
"Purple Haze",
"The Jimi Hendrix Experience"
],
"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:\nJimi Hendrix is the greatest guitarist of all time.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Purple Haze",
"The Jimi Hendrix Experience"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 490,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
69166
|
On the list of Vice Chairs for the North District Council, four of the Vice Chairmen have been affiliated with Heung Yee Kuk.
|
{
"evidence": [
"Heung Yee Kuk (constituency)_sentence_0",
"Heung Yee Kuk (constituency)_sentence_1"
],
"label": "NOT ENOUGH INFO"
}
|
feverous
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nOn the list of Vice Chairs for the North District Council, four of the Vice Chairmen have been affiliated with Heung Yee Kuk.\n",
"role": "user"
}
] |
Search terms not in claim
|
{
"ground_truth": {
"evidence": [
"Heung Yee Kuk (constituency)_sentence_0",
"Heung Yee Kuk (constituency)_sentence_1"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 491,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.229",
"value": "?search="
},
{
"operation": "search",
"time": "25.634",
"value": "North District Council Heung Yee Kuk"
},
{
"operation": "Now on",
"time": "28.185",
"value": "undefined"
},
{
"operation": "hyperlink",
"time": "216.113",
"value": "Cheung Yan-lung"
},
{
"operation": "Now on",
"time": "216.882",
"value": "Cheung Yan-lung"
},
{
"operation": "search",
"time": "242.481",
"value": "list of vice chairs for the North district council"
},
{
"operation": "Now on",
"time": "246.561",
"value": "undefined"
},
{
"operation": "search",
"time": "268.609",
"value": "Heung Yee Kuk"
},
{
"operation": "Now on",
"time": "270.804",
"value": "Heung Yee Kuk"
},
{
"operation": "hyperlink",
"time": "354.944",
"value": "Heung Yee Kuk (constituency)"
},
{
"operation": "Now on",
"time": "359.956",
"value": "Heung Yee Kuk (constituency)"
},
{
"operation": "Highlighting",
"time": "386.481",
"value": "Heung Yee Kuk (constituency)_sentence_0"
},
{
"operation": "Highlighting",
"time": "399.528",
"value": "Heung Yee Kuk (constituency)_sentence_1"
},
{
"operation": "search",
"time": "424.528",
"value": "North District Council china"
},
{
"operation": "Now on",
"time": "429.357",
"value": "undefined"
},
{
"operation": "finish",
"time": "445.32",
"value": "finish"
}
]
}
|
38649
|
Türkmennebit is the state-owned oil and gas company from Kazakhstan.
|
{
"evidence": [
"Türkmennebit_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:\nTürkmennebit is the state-owned oil and gas company from Kazakhstan.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Türkmennebit_sentence_0"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 492,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.64",
"value": "?search="
},
{
"operation": "search",
"time": "9.523",
"value": "Türkmennebit"
},
{
"operation": "Now on",
"time": "10.436",
"value": "Türkmennebit"
},
{
"operation": "Highlighting",
"time": "130.88",
"value": "Türkmennebit_sentence_0"
},
{
"operation": "finish",
"time": "142.555",
"value": "finish"
}
]
}
|
exfever_4170
|
Caitlyn Marie Jenner is an American media personality who won a silver medal in the decathlon in the leading international sporting events featuring summer and winter sports competitions, which are normally held every four years, alternating between the Winter and Games of the Olympiad, a major international multi-sport event normally held once every four years.
|
{
"evidence": [
"Caitlyn Jenner",
"Olympic Games",
"Summer Olympic Games"
],
"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:\nCaitlyn Marie Jenner is an American media personality who won a silver medal in the decathlon in the leading international sporting events featuring summer and winter sports competitions, which are normally held every four years, alternating between the Winter and Games of the Olympiad, a major international multi-sport event normally held once every four years.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Caitlyn Jenner",
"Olympic Games",
"Summer Olympic Games"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 493,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_6091
|
Jessica Phyllis Lange is the wealthiest American actress.
|
{
"evidence": [
"Jessica Lange",
"Triple Crown of Acting"
],
"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:\nJessica Phyllis Lange is the wealthiest American actress.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Jessica Lange",
"Triple Crown of Acting"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 494,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_7470
|
Judy Greer first rose to prominence in the late 1990s to early 2000s, appearing in a romantic fantasy comedy film directed by an Australian filmmaker.
|
{
"evidence": [
"Judy Greer",
"What Women Want",
"Nancy Meyers"
],
"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:\nJudy Greer first rose to prominence in the late 1990s to early 2000s, appearing in a romantic fantasy comedy film directed by an Australian filmmaker.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Judy Greer",
"What Women Want",
"Nancy Meyers"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 495,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_7684
|
Evan Goldberg is an English director, and one of his films is the 2013 American apocalyptic film, whose genre belongs to a category of film that emphasizes humor.
|
{
"evidence": [
"Evan Goldberg",
"This Is the End",
"Comedy 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:\nEvan Goldberg is an English director, and one of his films is the 2013 American apocalyptic film, whose genre belongs to a category of film that emphasizes humor.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Evan Goldberg",
"This Is the End",
"Comedy film"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 496,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_7475
|
Kevin Wayne Durant is a professional soccer player for the Brooklyn Nets in one of the major professional sports leagues in the United States and Canada.
|
{
"evidence": [
"Kevin Durant",
"National Basketball Association"
],
"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:\nKevin Wayne Durant is a professional soccer player for the Brooklyn Nets in one of the major professional sports leagues in the United States and Canada.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Kevin Durant",
"National Basketball Association"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 497,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
92371
|
The Punch and Judy Murders is a mystery novel by the American writer John Dickson Carr (1906–1977), who published it under the name of Carter Dickson that is about Kenwood Blake who is scheduled to marry marry his fellow British Secret Service operative, 24 hours from now but was sidetracked by an urgent telegram from Sir Henry asking him to come to Torquay to play an undercover role.
|
{
"evidence": [
"The Punch and Judy Murders_sentence_2",
"The Punch and Judy Murders_sentence_4",
"The Punch and Judy Murders_sentence_5"
],
"label": "SUPPORT"
}
|
feverous
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nThe Punch and Judy Murders is a mystery novel by the American writer John Dickson Carr (1906–1977), who published it under the name of Carter Dickson that is about Kenwood Blake who is scheduled to marry marry his fellow British Secret Service operative, 24 hours from now but was sidetracked by an urgent telegram from Sir Henry asking him to come to Torquay to play an undercover role.\n",
"role": "user"
}
] |
Multi-hop Reasoning
|
{
"ground_truth": {
"evidence": [
"The Punch and Judy Murders_sentence_2",
"The Punch and Judy Murders_sentence_4",
"The Punch and Judy Murders_sentence_5"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 498,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.274",
"value": "?search="
},
{
"operation": "search",
"time": "18.017",
"value": "John Dickson Carr"
},
{
"operation": "Now on",
"time": "19.619",
"value": "John Dickson Carr"
},
{
"operation": "Highlighting",
"time": "24.397",
"value": "John Dickson Carr_sentence_0"
},
{
"operation": "Highlighting",
"time": "24.399",
"value": "John Dickson Carr_title"
},
{
"operation": "search",
"time": "39.941",
"value": "Contains...Punch and Judy Murders"
},
{
"operation": "Now on",
"time": "42.386",
"value": "undefined"
},
{
"operation": "hyperlink",
"time": "45.314",
"value": "The Punch and Judy Murders"
},
{
"operation": "Now on",
"time": "45.88",
"value": "The Punch and Judy Murders"
},
{
"operation": "Highlighting",
"time": "53.626",
"value": "The Punch and Judy Murders_sentence_2"
},
{
"operation": "Page search",
"time": "78.496",
"value": "marry"
},
{
"operation": "Highlighting",
"time": "86.363",
"value": "The Punch and Judy Murders_sentence_4"
},
{
"operation": "Highlighting",
"time": "96.747",
"value": "The Punch and Judy Murders_sentence_5"
},
{
"operation": "finish",
"time": "139.492",
"value": "finish"
}
]
}
|
9060
|
Liver X receptor alpha is one of five liver X receptors, which activate thyroid hormones and play a role in regulating the suppression of target genes.
|
{
"evidence": [
"Liver X receptor_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:\nLiver X receptor alpha is one of five liver X receptors, which activate thyroid hormones and play a role in regulating the suppression of target genes.\n",
"role": "user"
}
] |
Numerical Reasoning
|
{
"ground_truth": {
"evidence": [
"Liver X receptor_sentence_3"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 499,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.79",
"value": "?search="
},
{
"operation": "search",
"time": "7.87",
"value": "Liver X receptor alpha"
},
{
"operation": "Now on",
"time": "8.398",
"value": "Liver X receptor alpha"
},
{
"operation": "hyperlink",
"time": "63.3",
"value": "liver X receptors"
},
{
"operation": "Now on",
"time": "63.881",
"value": "Liver X receptor"
},
{
"operation": "Page search",
"time": "70.746",
"value": "Liver X receptor alpha"
},
{
"operation": "Page search",
"time": "109.384",
"value": "alpha"
},
{
"operation": "back-button-clicked",
"time": "148.358",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "148.42",
"value": "Liver X receptor alpha"
},
{
"operation": "hyperlink",
"time": "169.136",
"value": "liver X receptors"
},
{
"operation": "Now on",
"time": "169.453",
"value": "Liver X receptor"
},
{
"operation": "back-button-clicked",
"time": "189.239",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "189.293",
"value": "Liver X receptor alpha"
},
{
"operation": "Page search",
"time": "202.597",
"value": "five"
},
{
"operation": "hyperlink",
"time": "289.615",
"value": "liver X receptors"
},
{
"operation": "Now on",
"time": "290.218",
"value": "Liver X receptor"
},
{
"operation": "Highlighting",
"time": "299.373",
"value": "Liver X receptor_sentence_3"
},
{
"operation": "finish",
"time": "327.67",
"value": "finish"
}
]
}
|
79488
|
During the Great Depression in 1933, the Kroon was overvalued and was exchanged for the Soviet ruble at a rate of 1 ruble = 0.8 kroon. .
|
{
"evidence": [
"Estonian kroon_sentence_14"
],
"label": "REFUTE"
}
|
feverous
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nDuring the Great Depression in 1933, the Kroon was overvalued and was exchanged for the Soviet ruble at a rate of 1 ruble = 0.8 kroon. .\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Estonian kroon_sentence_14"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 500,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.706",
"value": "undefined"
},
{
"operation": "search",
"time": "28.301",
"value": "Contains...the Kroon"
},
{
"operation": "Now on",
"time": "29.314",
"value": "undefined"
},
{
"operation": "search",
"time": "47.968",
"value": "Kroon"
},
{
"operation": "Now on",
"time": "48.807",
"value": "Kroon"
},
{
"operation": "hyperlink",
"time": "62.327",
"value": "Estonian kroon"
},
{
"operation": "Now on",
"time": "63.112",
"value": "Estonian kroon"
},
{
"operation": "Page search",
"time": "69.611",
"value": "depression"
},
{
"operation": "Highlighting",
"time": "75.699",
"value": "Estonian kroon_sentence_14"
},
{
"operation": "Highlighting",
"time": "152.334",
"value": "Estonian kroon_sentence_16"
},
{
"operation": "Highlighting deleted",
"time": "185.195",
"value": "Estonian kroon_sentence_16"
},
{
"operation": "finish",
"time": "198.332",
"value": "finish"
}
]
}
|
exfever_5002
|
Joseph McCarthy was an American lawyer and politician who served from January 1965 until June 1968 as the U.S. Senator from the state in the northeastern United States.
|
{
"evidence": [
"Robert F. Kennedy",
"New York (state)"
],
"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:\nJoseph McCarthy was an American lawyer and politician who served from January 1965 until June 1968 as the U.S. Senator from the state in the northeastern United States.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Robert F. Kennedy",
"New York (state)"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 501,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
37021
|
Physician Herbert Davies was elected as assistant-physician to the London Hospital in 1854, but was never elected as a physician there.
|
{
"evidence": [
"Herbert Davies_sentence_6",
"Herbert Davies_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:\nPhysician Herbert Davies was elected as assistant-physician to the London Hospital in 1854, but was never elected as a physician there.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Herbert Davies_sentence_6",
"Herbert Davies_sentence_5"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 502,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.508",
"value": "undefined"
},
{
"operation": "search",
"time": "7.835",
"value": "Herbert Davies"
},
{
"operation": "Now on",
"time": "8.663",
"value": "Herbert Davies"
},
{
"operation": "Highlighting",
"time": "22.644",
"value": "Herbert Davies_sentence_0"
},
{
"operation": "Highlighting",
"time": "38.013",
"value": "Herbert Davies_sentence_6"
},
{
"operation": "Highlighting deleted",
"time": "45.482",
"value": "Herbert Davies_sentence_0"
},
{
"operation": "Highlighting",
"time": "82.581",
"value": "Herbert Davies_sentence_5"
},
{
"operation": "finish",
"time": "95.073",
"value": "finish"
}
]
}
|
exfever_3310
|
From the Earth to the Moon is a 12-part HBO television miniseries that tells the story of the Apollo program in the 1960s and early 1970s and was co-produced by Ron Howard, Tom Hanks, Michael Bostick, and the American film and television producer and writer who co-founded Imagine Entertainment in 1986.
|
{
"evidence": [
"From the Earth to the Moon (miniseries)",
"Brian Grazer"
],
"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:\nFrom the Earth to the Moon is a 12-part HBO television miniseries that tells the story of the Apollo program in the 1960s and early 1970s and was co-produced by Ron Howard, Tom Hanks, Michael Bostick, and the American film and television producer and writer who co-founded Imagine Entertainment in 1986.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"From the Earth to the Moon (miniseries)",
"Brian Grazer"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 503,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_3252
|
Glee is a comedy-drama television series whose genre in which songs by the characters are interwoven into the narrative, sometimes accompanied by singing and dancing.
|
{
"evidence": [
"Glee (TV series)",
"Musical 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:\nGlee is a comedy-drama television series whose genre in which songs by the characters are interwoven into the narrative, sometimes accompanied by singing and dancing.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Glee (TV series)",
"Musical film"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 504,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
14433
|
According to the IEA, a policy advisor focusing on recycling, economic development, and environmental protection, energy security and renewable technology could make a direct contribution to heating.
|
{
"evidence": [
"International Energy Agency_sentence_77",
"International Energy Agency_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:\nAccording to the IEA, a policy advisor focusing on recycling, economic development, and environmental protection, energy security and renewable technology could make a direct contribution to heating.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"International Energy Agency_sentence_77",
"International Energy Agency_sentence_3"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 505,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.658",
"value": "?search="
},
{
"operation": "search",
"time": "12.483",
"value": "IEA"
},
{
"operation": "Now on",
"time": "14.373",
"value": "IEA"
},
{
"operation": "hyperlink",
"time": "54.454",
"value": "International Energy Agency"
},
{
"operation": "Now on",
"time": "56.444",
"value": "International Energy Agency"
},
{
"operation": "back-button-clicked",
"time": "61.82",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "62.056",
"value": "IEA"
},
{
"operation": "hyperlink",
"time": "65.711",
"value": "International Energy Agency"
},
{
"operation": "Now on",
"time": "66.664",
"value": "International Energy Agency"
},
{
"operation": "Page search",
"time": "93.05",
"value": "advisor"
},
{
"operation": "Highlighting",
"time": "115.615",
"value": "International Energy Agency_sentence_77"
},
{
"operation": "Page search",
"time": "180.91",
"value": "recycle"
},
{
"operation": "Page search",
"time": "195.991",
"value": "recycl"
},
{
"operation": "Page search",
"time": "212.62",
"value": "focus"
},
{
"operation": "Highlighting",
"time": "225.982",
"value": "International Energy Agency_sentence_3"
},
{
"operation": "Page search",
"time": "253.012",
"value": "heat"
},
{
"operation": "finish",
"time": "314.057",
"value": "finish"
}
]
}
|
50532
|
Eleuthera is an archipelago that incorporates the Harbour Island, it is adjacent to the North Atlantic Ocean in Barbados whose capital is Nassau.
|
{
"evidence": [
"Eleuthera_sentence_3",
"Barbados_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:\nEleuthera is an archipelago that incorporates the Harbour Island, it is adjacent to the North Atlantic Ocean in Barbados whose capital is Nassau.\n",
"role": "user"
}
] |
Multi-hop Reasoning
|
{
"ground_truth": {
"evidence": [
"Eleuthera_sentence_3",
"Barbados_sentence_3"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 506,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.414",
"value": "?search="
},
{
"operation": "search",
"time": "9.972",
"value": "Eleuthera"
},
{
"operation": "Now on",
"time": "11.243",
"value": "Eleuthera"
},
{
"operation": "Highlighting",
"time": "15.447",
"value": "Eleuthera_sentence_3"
},
{
"operation": "Highlighting",
"time": "25.535",
"value": "Eleuthera_sentence_5"
},
{
"operation": "Highlighting deleted",
"time": "27.888",
"value": "Eleuthera_sentence_5"
},
{
"operation": "search",
"time": "41.252",
"value": "Archipelago"
},
{
"operation": "Now on",
"time": "42.188",
"value": "Archipelago"
},
{
"operation": "Highlighting",
"time": "46.92",
"value": "Archipelago_sentence_4"
},
{
"operation": "Highlighting",
"time": "46.921",
"value": "Archipelago_title"
},
{
"operation": "back-button-clicked",
"time": "48.64",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "49.005",
"value": "Eleuthera"
},
{
"operation": "Highlighting",
"time": "52.008",
"value": "Eleuthera_sentence_5"
},
{
"operation": "Highlighting",
"time": "72.185",
"value": "Eleuthera_cell_0_3_1"
},
{
"operation": "Highlighting",
"time": "72.201",
"value": "Eleuthera_header_cell_0_3_0"
},
{
"operation": "Highlighting",
"time": "72.202",
"value": "Eleuthera_header_cell_0_0_0"
},
{
"operation": "Page search",
"time": "94.565",
"value": "Barbados"
},
{
"operation": "search",
"time": "121.974",
"value": "Barbados"
},
{
"operation": "Now on",
"time": "127.055",
"value": "Barbados"
},
{
"operation": "Highlighting",
"time": "137.045",
"value": "Barbados_sentence_3"
},
{
"operation": "back-button-clicked",
"time": "141.379",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "141.775",
"value": "Eleuthera"
},
{
"operation": "Highlighting",
"time": "144.54",
"value": "Eleuthera_sentence_3"
},
{
"operation": "back-button-clicked",
"time": "148.387",
"value": "back-button-clicked"
},
{
"operation": "back-button-clicked",
"time": "148.393",
"value": "back-button-clicked"
},
{
"operation": "back-button-clicked",
"time": "148.405",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "148.634",
"value": "?search="
},
{
"operation": "search",
"time": "154.537",
"value": "Barbados"
},
{
"operation": "Now on",
"time": "156.115",
"value": "Barbados"
},
{
"operation": "Highlighting",
"time": "160.356",
"value": "Barbados_sentence_3"
},
{
"operation": "finish",
"time": "180.773",
"value": "finish"
}
]
}
|
7435
|
Vytshegdosuchus, an extinct genus of paracrocodylomorph archosaur named in 1801 is known from a right ilium and was discovered from the Parotossuchus.
|
{
"evidence": [
"Vytshegdosuchus_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:\nVytshegdosuchus, an extinct genus of paracrocodylomorph archosaur named in 1801 is known from a right ilium and was discovered from the Parotossuchus.\n",
"role": "user"
}
] |
Numerical Reasoning
|
{
"ground_truth": {
"evidence": [
"Vytshegdosuchus_sentence_2"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 507,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.887",
"value": "?search="
},
{
"operation": "search",
"time": "31.105",
"value": "Vytshegdosuchus"
},
{
"operation": "Now on",
"time": "32.047",
"value": "Vytshegdosuchus"
},
{
"operation": "Highlighting",
"time": "36.047",
"value": "Vytshegdosuchus_sentence_0"
},
{
"operation": "Highlighting",
"time": "53.213",
"value": "Vytshegdosuchus_sentence_2"
},
{
"operation": "Highlighting deleted",
"time": "54.679",
"value": "Vytshegdosuchus_sentence_0"
},
{
"operation": "finish",
"time": "66.666",
"value": "finish"
}
]
}
|
37132
|
DUSP1 suppresses the activation of MAP kinase and specifies a protein with structural features similar to members of the non-receptor-type protein-tyrosine phosphorylases family.
|
{
"evidence": [
"DUSP1_sentence_4",
"DUSP1_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:\nDUSP1 suppresses the activation of MAP kinase and specifies a protein with structural features similar to members of the non-receptor-type protein-tyrosine phosphorylases family.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"DUSP1_sentence_4",
"DUSP1_sentence_2"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 508,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.067",
"value": "?search="
},
{
"operation": "search",
"time": "10.831",
"value": "DUSP1"
},
{
"operation": "Now on",
"time": "11.741",
"value": "DUSP1"
},
{
"operation": "Highlighting",
"time": "32.787",
"value": "DUSP1_sentence_4"
},
{
"operation": "Highlighting",
"time": "40.216",
"value": "DUSP1_sentence_2"
},
{
"operation": "finish",
"time": "58.358",
"value": "finish"
}
]
}
|
71696
|
Jairus Lyles (born July 6, 1995) is a 1.88 m Slovenian professional basketball player who played for the Slovenian League and the Salt Lake City Stars.
|
{
"evidence": [
"Jairus Lyles_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:\nJairus Lyles (born July 6, 1995) is a 1.88 m Slovenian professional basketball player who played for the Slovenian League and the Salt Lake City Stars.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Jairus Lyles_sentence_0"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 509,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.104",
"value": "?search="
},
{
"operation": "search",
"time": "13.239",
"value": "Jairus Lyles"
},
{
"operation": "Now on",
"time": "14.205",
"value": "Jairus Lyles"
},
{
"operation": "Highlighting",
"time": "17.409",
"value": "Jairus Lyles_sentence_0"
},
{
"operation": "finish",
"time": "31.642",
"value": "finish"
}
]
}
|
87531
|
Lucien Le Foyer (1872-1952), a French lawyer, was an enlightened Democrat, who ran unsuccessfully for reelection in 1910.
|
{
"evidence": [
"Lucien Le Foyer_sentence_0",
"Lucien Le Foyer_sentence_15",
"Lucien Le Foyer_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:\nLucien Le Foyer (1872-1952), a French lawyer, was an enlightened Democrat, who ran unsuccessfully for reelection in 1910.\n",
"role": "user"
}
] |
Multi-hop Reasoning
|
{
"ground_truth": {
"evidence": [
"Lucien Le Foyer_sentence_0",
"Lucien Le Foyer_sentence_15",
"Lucien Le Foyer_sentence_14"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 510,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.907",
"value": "?search="
},
{
"operation": "search",
"time": "10.799",
"value": "Lucien Le Foyer"
},
{
"operation": "Now on",
"time": "11.622",
"value": "Lucien Le Foyer"
},
{
"operation": "Highlighting",
"time": "26.951",
"value": "Lucien Le Foyer_sentence_0"
},
{
"operation": "Page search",
"time": "40.578",
"value": "enlightened"
},
{
"operation": "Highlighting",
"time": "43.524",
"value": "Lucien Le Foyer_sentence_15"
},
{
"operation": "Highlighting",
"time": "47.675",
"value": "Lucien Le Foyer_sentence_14"
},
{
"operation": "finish",
"time": "56.817",
"value": "finish"
}
]
}
|
exfever_8977
|
House of 1000 Corpses is a 2004 comedy horror film directed by an American singer, filmmaker, and voice actor.
|
{
"evidence": [
"House of 1000 Corpses",
"Rob Zombie"
],
"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:\nHouse of 1000 Corpses is a 2004 comedy horror film directed by an American singer, filmmaker, and voice actor.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"House of 1000 Corpses",
"Rob Zombie"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 511,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
15370
|
In the novel Five Smooth Stones, David comes back from serving as a veterinarian and tragedy ensues.
|
{
"evidence": [
"Five Smooth Stones_sentence_6",
"Five Smooth Stones_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:\nIn the novel Five Smooth Stones, David comes back from serving as a veterinarian and tragedy ensues.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Five Smooth Stones_sentence_6",
"Five Smooth Stones_sentence_5"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 512,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.669",
"value": "?search="
},
{
"operation": "search",
"time": "8.631",
"value": "Five Smooth Stones"
},
{
"operation": "Now on",
"time": "9.482",
"value": "Five Smooth Stones"
},
{
"operation": "Page search",
"time": "15.44",
"value": "comes back"
},
{
"operation": "Highlighting",
"time": "26.988",
"value": "Five Smooth Stones_sentence_6"
},
{
"operation": "Highlighting",
"time": "48.995",
"value": "Five Smooth Stones_sentence_5"
},
{
"operation": "Highlighting",
"time": "57.581",
"value": "Five Smooth Stones_sentence_7"
},
{
"operation": "Highlighting deleted",
"time": "73.493",
"value": "Five Smooth Stones_sentence_7"
},
{
"operation": "finish",
"time": "82.821",
"value": "finish"
}
]
}
|
85734
|
The Philactis plant, not related to the the sunflower tribe, is in the asteraceae family and the plantae kingdom.
|
{
"evidence": [
"Philactis_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 Philactis plant, not related to the the sunflower tribe, is in the asteraceae family and the plantae kingdom.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Philactis_sentence_0"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 513,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.509",
"value": "?search="
},
{
"operation": "search",
"time": "61.219",
"value": "The Philactis plant"
},
{
"operation": "Now on",
"time": "62.176",
"value": "undefined"
},
{
"operation": "hyperlink",
"time": "70.759",
"value": "Philactis"
},
{
"operation": "Now on",
"time": "71.475",
"value": "Philactis"
},
{
"operation": "Highlighting",
"time": "77.085",
"value": "Philactis_sentence_0"
},
{
"operation": "finish",
"time": "105.003",
"value": "finish"
}
]
}
|
25921
|
Eugene Simon (born 11 June 1992), the son of Anton and Teresa Simon, only attended one school in his lifetime.
|
{
"evidence": [
"Eugene Simon_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:\nEugene Simon (born 11 June 1992), the son of Anton and Teresa Simon, only attended one school in his lifetime.\n",
"role": "user"
}
] |
Numerical Reasoning
|
{
"ground_truth": {
"evidence": [
"Eugene Simon_sentence_5"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 514,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.95",
"value": "?search="
},
{
"operation": "search",
"time": "12.765",
"value": "Eugene Simon"
},
{
"operation": "Now on",
"time": "13.749",
"value": "Eugene Simon"
},
{
"operation": "Highlighting",
"time": "38.045",
"value": "Eugene Simon_sentence_3"
},
{
"operation": "Highlighting",
"time": "41.365",
"value": "Eugene Simon_sentence_5"
},
{
"operation": "Highlighting deleted",
"time": "42.189",
"value": "Eugene Simon_sentence_3"
},
{
"operation": "finish",
"time": "53.764",
"value": "finish"
}
]
}
|
23087
|
Sweden's Jonas Sandqvist made the most appearances (26) in the Allsvenkan league, a Swedish professional league for men's association football clubs, among the 2007 Malmö FF season squad members.
|
{
"evidence": [
"Jonas Sandqvist_sentence_0",
"Malmö FF_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:\nSweden's Jonas Sandqvist made the most appearances (26) in the Allsvenkan league, a Swedish professional league for men's association football clubs, among the 2007 Malmö FF season squad members.\n",
"role": "user"
}
] |
Entity Disambiguation
|
{
"ground_truth": {
"evidence": [
"Jonas Sandqvist_sentence_0",
"Malmö FF_sentence_3"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 515,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.361",
"value": "?search="
},
{
"operation": "search",
"time": "30.153",
"value": "Contains...Sweden's Jonas Sandqvist"
},
{
"operation": "Now on",
"time": "31.76",
"value": "undefined"
},
{
"operation": "hyperlink",
"time": "40.527",
"value": "Jonas Sandqvist"
},
{
"operation": "Now on",
"time": "41.239",
"value": "Jonas Sandqvist"
},
{
"operation": "Highlighting",
"time": "49.211",
"value": "Jonas Sandqvist_sentence_0"
},
{
"operation": "Now on",
"time": "129.217",
"value": "Jonas Sandqvist"
},
{
"operation": "hyperlink",
"time": "277.509",
"value": "Malmö FF"
},
{
"operation": "Now on",
"time": "279.648",
"value": "Malmö FF"
},
{
"operation": "Highlighting",
"time": "287.978",
"value": "Malmö FF_sentence_3"
},
{
"operation": "search",
"time": "442.658",
"value": "Allsvenkan league"
},
{
"operation": "Now on",
"time": "444.285",
"value": "undefined"
},
{
"operation": "search",
"time": "485.688",
"value": "Allsvenkan"
},
{
"operation": "Now on",
"time": "486.297",
"value": "undefined"
},
{
"operation": "search",
"time": "540.741",
"value": "2007 Allsvenka"
},
{
"operation": "Now on",
"time": "542.077",
"value": "undefined"
},
{
"operation": "search",
"time": "577.948",
"value": "Jonas Sandqvist"
},
{
"operation": "Now on",
"time": "579.076",
"value": "Jonas Sandqvist"
},
{
"operation": "Now on",
"time": "704.286",
"value": "Jonas Sandqvist"
},
{
"operation": "finish",
"time": "713.52",
"value": "finish"
}
]
}
|
exfever_8969
|
Respiratory diseases are the most common diseases in Asia.
|
{
"evidence": [
"Respiratory disease",
"Trachea"
],
"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:\nRespiratory diseases are the most common diseases in Asia.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Respiratory disease",
"Trachea"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 516,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
74180
|
Laurel van der Wal is falsely believed to have worked on many projects and also focused on the design of manned spacecraft she was also the second woman appointed to the Los Angeles Board of Airport Commissioners (formed in 1928)
|
{
"evidence": [
"Laurel van der Wal_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:\nLaurel van der Wal is falsely believed to have worked on many projects and also focused on the design of manned spacecraft she was also the second woman appointed to the Los Angeles Board of Airport Commissioners (formed in 1928)\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Laurel van der Wal_sentence_17"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 517,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.883",
"value": "?search="
},
{
"operation": "search",
"time": "9.314",
"value": "Laurel van der Wal"
},
{
"operation": "Now on",
"time": "10.699",
"value": "Laurel van der Wal"
},
{
"operation": "Highlighting",
"time": "46.709",
"value": "Laurel van der Wal_sentence_17"
},
{
"operation": "finish",
"time": "57.498",
"value": "finish"
}
]
}
|
exfever_2569
|
Lauren Bacall was an English actress who married an American film and stage actor.
|
{
"evidence": [
"Lauren Bacall",
"Humphrey Bogart"
],
"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:\nLauren Bacall was an English actress who married an American film and stage actor.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Lauren Bacall",
"Humphrey Bogart"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 518,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
77620
|
Thomas Erdbrink (born January 27, 1976) speaks English, Persian, German and Dutch, and regularly appears on European and American media outlets, including a TV documentary series called Onze man in Teheran, which won the 2015 Zilveren Nipkowschijf.
|
{
"evidence": [
"Thomas Erdbrink_sentence_0",
"Thomas Erdbrink_sentence_5",
"Thomas Erdbrink_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:\nThomas Erdbrink (born January 27, 1976) speaks English, Persian, German and Dutch, and regularly appears on European and American media outlets, including a TV documentary series called Onze man in Teheran, which won the 2015 Zilveren Nipkowschijf.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Thomas Erdbrink_sentence_0",
"Thomas Erdbrink_sentence_5",
"Thomas Erdbrink_sentence_6"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 519,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.272",
"value": "?search="
},
{
"operation": "search",
"time": "20.177",
"value": "Thomas Erdbrink"
},
{
"operation": "Now on",
"time": "20.833",
"value": "Thomas Erdbrink"
},
{
"operation": "Highlighting",
"time": "34.985",
"value": "Thomas Erdbrink_sentence_0"
},
{
"operation": "Highlighting",
"time": "57.351",
"value": "Thomas Erdbrink_sentence_5"
},
{
"operation": "Highlighting",
"time": "72.103",
"value": "Thomas Erdbrink_sentence_6"
},
{
"operation": "finish",
"time": "97.174",
"value": "finish"
}
]
}
|
46924
|
Ahoada West is a Local Government Area of Rivers State, Nigeria that was extracted in 1996 from the old Ahoada Local Government and comprised of the Ekpeyes (Ubies and Ibuduyas), Engenni and Ogbogolo communities.
|
{
"evidence": [
"Ahoada West_sentence_0",
"Ahoada West_sentence_1",
"Ahoada West_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:\nAhoada West is a Local Government Area of Rivers State, Nigeria that was extracted in 1996 from the old Ahoada Local Government and comprised of the Ekpeyes (Ubies and Ibuduyas), Engenni and Ogbogolo communities.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Ahoada West_sentence_0",
"Ahoada West_sentence_1",
"Ahoada West_sentence_3"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 520,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.701",
"value": "?search="
},
{
"operation": "search",
"time": "13.461",
"value": "Ahoada West"
},
{
"operation": "Now on",
"time": "14.357",
"value": "Ahoada West"
},
{
"operation": "Highlighting",
"time": "21.422",
"value": "Ahoada West_sentence_0"
},
{
"operation": "Highlighting",
"time": "29.473",
"value": "Ahoada West_sentence_1"
},
{
"operation": "Highlighting",
"time": "42.585",
"value": "Ahoada West_sentence_3"
},
{
"operation": "finish",
"time": "57.735",
"value": "finish"
}
]
}
|
23710
|
From 1995-1999 Gao Xiqing served as the General Counsel and the Director of Public Offerings of Singapore Securities Regulatory Commission (SSRC) and currently serves on the Duke University Board of Trustees.
|
{
"evidence": [
"Gao Xiqing_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:\nFrom 1995-1999 Gao Xiqing served as the General Counsel and the Director of Public Offerings of Singapore Securities Regulatory Commission (SSRC) and currently serves on the Duke University Board of Trustees.\n",
"role": "user"
}
] |
Numerical Reasoning
|
{
"ground_truth": {
"evidence": [
"Gao Xiqing_sentence_4"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 521,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.26",
"value": "undefined"
},
{
"operation": "search",
"time": "17.216",
"value": "Gao Xiqing"
},
{
"operation": "Now on",
"time": "18.061",
"value": "Gao Xiqing"
},
{
"operation": "Page search",
"time": "28.888",
"value": "general"
},
{
"operation": "Highlighting",
"time": "42.129",
"value": "Gao Xiqing_sentence_4"
},
{
"operation": "finish",
"time": "184.978",
"value": "finish"
}
]
}
|
exfever_5192
|
British Columbia is the only province in Canada with rugged landscapes.
|
{
"evidence": [
"British Columbia",
"Pacific Ocean",
"Land and water hemispheres"
],
"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:\nBritish Columbia is the only province in Canada with rugged landscapes.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"British Columbia",
"Pacific Ocean",
"Land and water hemispheres"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 522,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_6147
|
William Hanna was the sixteenth American animator to win eight Emmy Awards.?
|
{
"evidence": [
"William Hanna",
"Hanna-Barbera"
],
"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:\nWilliam Hanna was the sixteenth American animator to win eight Emmy Awards.?\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"William Hanna",
"Hanna-Barbera"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 523,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_6057
|
The Hermetic Order of the Golden Dawn, a secret society devoted to the study of occult, metaphysics, and paranormal activities, is known as an organization created for the practice of other forms of occult magic.
|
{
"evidence": [
"Hermetic Order of the Golden Dawn",
"Magical organization"
],
"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 Hermetic Order of the Golden Dawn, a secret society devoted to the study of occult, metaphysics, and paranormal activities, is known as an organization created for the practice of other forms of occult magic.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Hermetic Order of the Golden Dawn",
"Magical organization"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 524,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_2354
|
JAG is an American television series with a U.S. Navy theme that focuses on a genre of film and television that generally focuses on narratives regarding legal practice and the justice system.
|
{
"evidence": [
"JAG (TV series)",
"Legal drama"
],
"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:\nJAG is an American television series with a U.S. Navy theme that focuses on a genre of film and television that generally focuses on narratives regarding legal practice and the justice system.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"JAG (TV series)",
"Legal drama"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 525,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
88055
|
There are fourteen Twenty20 International cricketers, a type of cricket that has a maximum of twenty overs.
|
{
"evidence": [
"List of Twenty20 International cricketers_sentence_1",
"List of Twenty20 International cricketers_sentence_3",
"List of Twenty20 International cricketers_sentence_6"
],
"label": "REFUTE"
}
|
feverous
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nThere are fourteen Twenty20 International cricketers, a type of cricket that has a maximum of twenty overs.\n",
"role": "user"
}
] |
Numerical Reasoning
|
{
"ground_truth": {
"evidence": [
"List of Twenty20 International cricketers_sentence_1",
"List of Twenty20 International cricketers_sentence_3",
"List of Twenty20 International cricketers_sentence_6"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 526,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.954",
"value": "?search="
},
{
"operation": "search",
"time": "15.098",
"value": "Twenty20 International cricket"
},
{
"operation": "Now on",
"time": "16.283",
"value": "Twenty20 International"
},
{
"operation": "Highlighting",
"time": "23.716",
"value": "Twenty20 International_sentence_2"
},
{
"operation": "Highlighting",
"time": "23.717",
"value": "Twenty20 International_title"
},
{
"operation": "search",
"time": "155.824",
"value": "Contains...Twenty20 International cricketers"
},
{
"operation": "Now on",
"time": "157.743",
"value": "undefined"
},
{
"operation": "search",
"time": "198.352",
"value": "List of Twenty20 International cricketers"
},
{
"operation": "Now on",
"time": "200.305",
"value": "List of Twenty20 International cricketers"
},
{
"operation": "Highlighting",
"time": "236.835",
"value": "List of Twenty20 International cricketers_sentence_1"
},
{
"operation": "Highlighting",
"time": "242.928",
"value": "List of Twenty20 International cricketers_sentence_3"
},
{
"operation": "Highlighting",
"time": "247.166",
"value": "List of Twenty20 International cricketers_sentence_6"
},
{
"operation": "finish",
"time": "260.356",
"value": "finish"
}
]
}
|
52676
|
The Valois station closes at 10PM everyday.
|
{
"evidence": [
"Valois station_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:\nThe Valois station closes at 10PM everyday.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Valois station_sentence_2"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 527,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.785",
"value": "?search="
},
{
"operation": "search",
"time": "25.615",
"value": "valois station"
},
{
"operation": "Now on",
"time": "26.759",
"value": "Valois station"
},
{
"operation": "hyperlink",
"time": "118.012",
"value": "Vaudreuil–Hudson line"
},
{
"operation": "Now on",
"time": "119.156",
"value": "Vaudreuil–Hudson line"
},
{
"operation": "search",
"time": "152.827",
"value": "\"valois station\""
},
{
"operation": "Now on",
"time": "153.582",
"value": "Valois station"
},
{
"operation": "search",
"time": "212.629",
"value": "insource"
},
{
"operation": "Now on",
"time": "215.226",
"value": "undefined"
},
{
"operation": "search",
"time": "310.193",
"value": "the valois station"
},
{
"operation": "Now on",
"time": "312.151",
"value": "undefined"
},
{
"operation": "hyperlink",
"time": "324.627",
"value": "Crépy-en-Valois station"
},
{
"operation": "Now on",
"time": "325.444",
"value": "Crépy-en-Valois station"
},
{
"operation": "back-button-clicked",
"time": "344.66",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "345.959",
"value": "undefined"
},
{
"operation": "hyperlink",
"time": "355.708",
"value": "Valois"
},
{
"operation": "Now on",
"time": "375.601",
"value": "Valois, Pointe-Claire"
},
{
"operation": "back-button-clicked",
"time": "382.644",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "383.814",
"value": "Valois"
},
{
"operation": "hyperlink",
"time": "385.059",
"value": "Valois station"
},
{
"operation": "Now on",
"time": "385.447",
"value": "Valois station"
},
{
"operation": "hyperlink",
"time": "427.915",
"value": "Vaudreuil–Hudson line"
},
{
"operation": "Now on",
"time": "428.606",
"value": "Vaudreuil–Hudson line"
},
{
"operation": "back-button-clicked",
"time": "471.276",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "471.437",
"value": "Valois station"
},
{
"operation": "Highlighting",
"time": "507.856",
"value": "Valois station_sentence_2"
},
{
"operation": "finish",
"time": "513.366",
"value": "finish"
}
]
}
|
exfever_263
|
Furious 7 is a 2014 American action film that stars an American actor who had his first starring role in John Singleton's coming-of-age hood film Baby Boy.
|
{
"evidence": [
"Furious 7",
"Tyrese Gibson"
],
"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:\nFurious 7 is a 2014 American action film that stars an American actor who had his first starring role in John Singleton's coming-of-age hood film Baby Boy.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Furious 7",
"Tyrese Gibson"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 528,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
94644
|
Sydney Rae White is the frontman of the band called The Wild Things.
|
{
"evidence": [
"Sydney Rae White_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:\nSydney Rae White is the frontman of the band called The Wild Things.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Sydney Rae White_sentence_3"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 529,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.283",
"value": "?search="
},
{
"operation": "search",
"time": "9.483",
"value": "Sydney Rae White"
},
{
"operation": "Now on",
"time": "10.572",
"value": "Sydney Rae White"
},
{
"operation": "Page search",
"time": "21.949",
"value": "Wild"
},
{
"operation": "Highlighting",
"time": "34.419",
"value": "Sydney Rae White_sentence_3"
},
{
"operation": "Highlighting",
"time": "38.948",
"value": "Sydney Rae White_sentence_5"
},
{
"operation": "Highlighting deleted",
"time": "42.828",
"value": "Sydney Rae White_sentence_5"
},
{
"operation": "Page search",
"time": "74.388",
"value": "frontman"
},
{
"operation": "search",
"time": "84.1",
"value": "The Wild Things"
},
{
"operation": "Now on",
"time": "85.536",
"value": "undefined"
},
{
"operation": "finish",
"time": "108.309",
"value": "finish"
}
]
}
|
82919
|
The Jenkins Orphanage was located at a National Historic Landmark, the Old Marine Hospital at 20 Franklin Street.
|
{
"evidence": [
"Jenkins Orphanage_sentence_4",
"Jenkins Orphanage_sentence_5"
],
"label": "SUPPORT"
}
|
feverous
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nThe Jenkins Orphanage was located at a National Historic Landmark, the Old Marine Hospital at 20 Franklin Street.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Jenkins Orphanage_sentence_4",
"Jenkins Orphanage_sentence_5"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 530,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.025",
"value": "?search="
},
{
"operation": "search",
"time": "11.027",
"value": "Jenkins Orphanage"
},
{
"operation": "Now on",
"time": "11.76",
"value": "Jenkins Orphanage"
},
{
"operation": "Highlighting",
"time": "349.785",
"value": "Jenkins Orphanage_sentence_0"
},
{
"operation": "Highlighting deleted",
"time": "349.913",
"value": "Jenkins Orphanage_sentence_0"
},
{
"operation": "Highlighting",
"time": "356.24",
"value": "Jenkins Orphanage_sentence_5"
},
{
"operation": "Highlighting",
"time": "363.322",
"value": "Jenkins Orphanage_sentence_4"
},
{
"operation": "Highlighting deleted",
"time": "365.851",
"value": "Jenkins Orphanage_sentence_5"
},
{
"operation": "Highlighting",
"time": "366.272",
"value": "Jenkins Orphanage_sentence_5"
},
{
"operation": "finish",
"time": "378.207",
"value": "finish"
}
]
}
|
exfever_728
|
J. Cole is an African rapper, singer, songwriter, and record producer who released an album release by Roc Nation.
|
{
"evidence": [
"J. Cole",
"Born Sinner"
],
"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:\nJ. Cole is an African rapper, singer, songwriter, and record producer who released an album release by Roc Nation.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"J. Cole",
"Born Sinner"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 531,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
35539
|
The aurelian is Paul Pilgram, his life is empty and he imagines to be in butterfly places such as France but was exiled in Berlin.
|
{
"evidence": [
"The Aurelian_sentence_3",
"The Aurelian_sentence_4",
"The Aurelian_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:\nThe aurelian is Paul Pilgram, his life is empty and he imagines to be in butterfly places such as France but was exiled in Berlin.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"The Aurelian_sentence_3",
"The Aurelian_sentence_4",
"The Aurelian_sentence_6"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 532,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.691",
"value": "?search="
},
{
"operation": "search",
"time": "18.604",
"value": "Contains...Paul Pilgram"
},
{
"operation": "Now on",
"time": "19.493",
"value": "undefined"
},
{
"operation": "hyperlink",
"time": "22.236",
"value": "The Aurelian"
},
{
"operation": "Now on",
"time": "22.745",
"value": "The Aurelian"
},
{
"operation": "Highlighting",
"time": "28.772",
"value": "The Aurelian_sentence_3"
},
{
"operation": "Highlighting",
"time": "31.74",
"value": "The Aurelian_sentence_4"
},
{
"operation": "Highlighting",
"time": "35.612",
"value": "The Aurelian_sentence_6"
},
{
"operation": "finish",
"time": "75.767",
"value": "finish"
}
]
}
|
45186
|
Brown Bear Seamount, part of the Cobb-Eickelberg Seamount chain which is a range of undersea mountains located in the Pacific Ocean, is separated in two areas, the Northwestern and Southern side.
|
{
"evidence": [
"Brown Bear Seamount_sentence_1",
"Cobb–Eickelberg Seamount chain_sentence_0",
"Cobb–Eickelberg Seamount chain_sentence_4",
"Brown Bear Seamount_sentence_4"
],
"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:\nBrown Bear Seamount, part of the Cobb-Eickelberg Seamount chain which is a range of undersea mountains located in the Pacific Ocean, is separated in two areas, the Northwestern and Southern side.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Brown Bear Seamount_sentence_1",
"Cobb–Eickelberg Seamount chain_sentence_0",
"Cobb–Eickelberg Seamount chain_sentence_4",
"Brown Bear Seamount_sentence_4"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 533,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.931",
"value": "?search="
},
{
"operation": "search",
"time": "8.792",
"value": "Brown Bear Seamount"
},
{
"operation": "Now on",
"time": "9.519",
"value": "Brown Bear Seamount"
},
{
"operation": "hyperlink",
"time": "45.487",
"value": "Bear Seamount"
},
{
"operation": "Now on",
"time": "46.243",
"value": "Bear Seamount"
},
{
"operation": "back-button-clicked",
"time": "76.211",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "76.313",
"value": "Brown Bear Seamount"
},
{
"operation": "Highlighting",
"time": "81.915",
"value": "Brown Bear Seamount_sentence_1"
},
{
"operation": "hyperlink",
"time": "92.245",
"value": "Cobb hotspot"
},
{
"operation": "Now on",
"time": "93.056",
"value": "Cobb hotspot"
},
{
"operation": "hyperlink",
"time": "102.921",
"value": "Cobb-Eicklberg seamount chain"
},
{
"operation": "Now on",
"time": "104.118",
"value": "Cobb–Eickelberg Seamount chain"
},
{
"operation": "Highlighting",
"time": "110.956",
"value": "Cobb–Eickelberg Seamount chain_sentence_0"
},
{
"operation": "Highlighting",
"time": "123.027",
"value": "Cobb–Eickelberg Seamount chain_sentence_4"
},
{
"operation": "hyperlink",
"time": "126.864",
"value": "Brown Bear"
},
{
"operation": "Now on",
"time": "127.275",
"value": "Brown Bear Seamount"
},
{
"operation": "hyperlink",
"time": "152.913",
"value": "ocean spreading"
},
{
"operation": "Now on",
"time": "155.129",
"value": "Mid-ocean ridge"
},
{
"operation": "back-button-clicked",
"time": "171.581",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "171.685",
"value": "Brown Bear Seamount"
},
{
"operation": "Highlighting",
"time": "173.781",
"value": "Brown Bear Seamount_sentence_4"
},
{
"operation": "finish",
"time": "189.511",
"value": "finish"
}
]
}
|
exfever_8184
|
Injustice 2 was initially released for the game consoles that include the home video game console developed by Microsoft.
|
{
"evidence": [
"Injustice 2",
"Xbox One"
],
"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:\nInjustice 2 was initially released for the game consoles that include the home video game console developed by Microsoft.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Injustice 2",
"Xbox One"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 534,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_1785
|
Lynyrd Skynyrd is the most sought-after band of the 1980s.
|
{
"evidence": [
"Lynyrd Skynyrd",
"Rock music"
],
"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:\nLynyrd Skynyrd is the most sought-after band of the 1980s.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Lynyrd Skynyrd",
"Rock music"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 535,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
57856
|
Jake Thompson pitched for the Lehigh Valley IronPigs in 2016 but failed to win the International League Most Valuable Pitcher Award.
|
{
"evidence": [
"Jake Thompson_sentence_24"
],
"label": "REFUTE"
}
|
feverous
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nJake Thompson pitched for the Lehigh Valley IronPigs in 2016 but failed to win the International League Most Valuable Pitcher Award.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Jake Thompson_sentence_24"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 536,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.043",
"value": "?search="
},
{
"operation": "search",
"time": "10.913",
"value": "Jake Thompson"
},
{
"operation": "Now on",
"time": "12.196",
"value": "Jake Thompson"
},
{
"operation": "Page search",
"time": "28.19",
"value": "Valley"
},
{
"operation": "Highlighting",
"time": "38.109",
"value": "Jake Thompson_sentence_22"
},
{
"operation": "Highlighting",
"time": "57.349",
"value": "Jake Thompson_sentence_24"
},
{
"operation": "Highlighting deleted",
"time": "60.053",
"value": "Jake Thompson_sentence_22"
},
{
"operation": "finish",
"time": "73.87",
"value": "finish"
}
]
}
|
31593
|
Zack Schilawski was an American former soccer player born in April 15, 1987 at Raleigh, North Carolina who played forward for the Wake Forest Demon Deacons from 2006 to 2009—wherein he was able to earn soccer awards—and Cary Clarets in 2007 to 2009 wherein he afterwards decided not to continue with football.
|
{
"evidence": [
"Zack Schilawski_sentence_0",
"Zack Schilawski_sentence_15",
"Zack Schilawski_sentence_14"
],
"label": "REFUTE"
}
|
feverous
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nZack Schilawski was an American former soccer player born in April 15, 1987 at Raleigh, North Carolina who played forward for the Wake Forest Demon Deacons from 2006 to 2009—wherein he was able to earn soccer awards—and Cary Clarets in 2007 to 2009 wherein he afterwards decided not to continue with football.\n",
"role": "user"
}
] |
Multi-hop Reasoning
|
{
"ground_truth": {
"evidence": [
"Zack Schilawski_sentence_0",
"Zack Schilawski_sentence_15",
"Zack Schilawski_sentence_14"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 537,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.883",
"value": "?search="
},
{
"operation": "search",
"time": "34.113",
"value": "zack schilawskii"
},
{
"operation": "Now on",
"time": "35.048",
"value": "undefined"
},
{
"operation": "hyperlink",
"time": "45.183",
"value": "zack schilawski"
},
{
"operation": "Now on",
"time": "46.157",
"value": "undefined"
},
{
"operation": "hyperlink",
"time": "49.38",
"value": "Zack Schilawski"
},
{
"operation": "Now on",
"time": "50.371",
"value": "Zack Schilawski"
},
{
"operation": "Highlighting",
"time": "61.387",
"value": "Zack Schilawski_sentence_0"
},
{
"operation": "Highlighting",
"time": "179.595",
"value": "Zack Schilawski_sentence_15"
},
{
"operation": "Highlighting",
"time": "189.32",
"value": "Zack Schilawski_sentence_14"
},
{
"operation": "Page search",
"time": "207.06",
"value": "football"
},
{
"operation": "finish",
"time": "257.786",
"value": "finish"
}
]
}
|
38080
|
In the 2011 census in Africa, there were 839 people.
|
{
"evidence": [
"South African National Census of 2011_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:\nIn the 2011 census in Africa, there were 839 people.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"South African National Census of 2011_sentence_0"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 538,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.571",
"value": "?search="
},
{
"operation": "search",
"time": "42.467",
"value": "2011 census africa"
},
{
"operation": "Now on",
"time": "44.214",
"value": "undefined"
},
{
"operation": "search",
"time": "50.281",
"value": "africa 2011 census"
},
{
"operation": "Now on",
"time": "51.275",
"value": "undefined"
},
{
"operation": "search",
"time": "61.319",
"value": "839 people 2011 census"
},
{
"operation": "Now on",
"time": "62.955",
"value": "undefined"
},
{
"operation": "search",
"time": "72.657",
"value": "africa census"
},
{
"operation": "Now on",
"time": "73.533",
"value": "undefined"
},
{
"operation": "hyperlink",
"time": "78.756",
"value": "Census in Botswana"
},
{
"operation": "Now on",
"time": "79.636",
"value": "Census in Botswana"
},
{
"operation": "hyperlink",
"time": "83.675",
"value": "Census"
},
{
"operation": "Now on",
"time": "84.793",
"value": "Census"
},
{
"operation": "back-button-clicked",
"time": "91.455",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "91.542",
"value": "Census in Botswana"
},
{
"operation": "search",
"time": "171",
"value": "census of africa"
},
{
"operation": "Now on",
"time": "172.933",
"value": "undefined"
},
{
"operation": "search",
"time": "194.386",
"value": "2011 census"
},
{
"operation": "Now on",
"time": "195.184",
"value": "2011 census"
},
{
"operation": "hyperlink",
"time": "250.069",
"value": "South African National Census of 2011"
},
{
"operation": "Now on",
"time": "251.349",
"value": "South African National Census of 2011"
},
{
"operation": "Page search",
"time": "263.099",
"value": "839"
},
{
"operation": "Highlighting",
"time": "277.985",
"value": "South African National Census of 2011_sentence_0"
},
{
"operation": "finish",
"time": "280.163",
"value": "finish"
}
]
}
|
exfever_1463
|
Oliver Reed was an English actor known for notable films such as the 1966 British-Canadian adventure western film directed by Sidney Hayers.
|
{
"evidence": [
"Oliver Reed",
"The Trap (1966 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:\nOliver Reed was an English actor known for notable films such as the 1966 British-Canadian adventure western film directed by Sidney Hayers.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Oliver Reed",
"The Trap (1966 film)"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 539,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_5652
|
The Incredible Hulk is a 2008 American superhero film directed by a Filipino film director and producer, whose films include the first two Transporter films.
|
{
"evidence": [
"The Incredible Hulk (film)",
"Louis Leterrier"
],
"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 Incredible Hulk is a 2008 American superhero film directed by a Filipino film director and producer, whose films include the first two Transporter films.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"The Incredible Hulk (film)",
"Louis Leterrier"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 540,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
64778
|
To recruit skilled staff for the project, a newspaper advertisement was posted; but most of the applicants had no records of prior education because those were lost in the Holocaust or during immigration.
|
{
"evidence": [
"WEIZAC_sentence_19",
"WEIZAC_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:\nTo recruit skilled staff for the project, a newspaper advertisement was posted; but most of the applicants had no records of prior education because those were lost in the Holocaust or during immigration.\n",
"role": "user"
}
] |
Search terms not in claim
|
{
"ground_truth": {
"evidence": [
"WEIZAC_sentence_19",
"WEIZAC_sentence_20"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 541,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.847",
"value": "?search="
},
{
"operation": "search",
"time": "54.513",
"value": "WEIZAC"
},
{
"operation": "Now on",
"time": "56.482",
"value": "WEIZAC"
},
{
"operation": "Page search",
"time": "90.34",
"value": "advert"
},
{
"operation": "Highlighting",
"time": "93.994",
"value": "WEIZAC_sentence_19"
},
{
"operation": "Highlighting",
"time": "99.794",
"value": "WEIZAC_sentence_20"
},
{
"operation": "finish",
"time": "113.302",
"value": "finish"
}
]
}
|
exfever_7246
|
The great white shark belongs to the largest order of sharks.
|
{
"evidence": [
"Great white shark",
"Marine mammal"
],
"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 great white shark belongs to the largest order of sharks.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Great white shark",
"Marine mammal"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 542,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_4624
|
Bungie was established in May 1992 by Alex Seropian who later brought in programmer who co-founded the video game studio
|
{
"evidence": [
"Bungie",
"Jason Jones (programmer)"
],
"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:\nBungie was established in May 1992 by Alex Seropian who later brought in programmer who co-founded the video game studio\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Bungie",
"Jason Jones (programmer)"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 543,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_7207
|
Thomas Chong is a Canadian-American actor known for his Cheech & Chong comedy albums whose theme is about drugs, which is a chemical substance that changes nervous system functions from the Cannabis plant.
|
{
"evidence": [
"Tommy Chong",
"Cannabis (drug)",
"Psychoactive drug"
],
"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:\nThomas Chong is a Canadian-American actor known for his Cheech & Chong comedy albums whose theme is about drugs, which is a chemical substance that changes nervous system functions from the Cannabis plant.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Tommy Chong",
"Cannabis (drug)",
"Psychoactive drug"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 544,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
53754
|
Dabnou, which is near the third largest town in Niger, has two districts headed by Elhadji Gado Nahantchi and Mourtala Aboubacar Nabirni.
|
{
"evidence": [
"Dabnou_sentence_1",
"Tahoua_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:\nDabnou, which is near the third largest town in Niger, has two districts headed by Elhadji Gado Nahantchi and Mourtala Aboubacar Nabirni.\n",
"role": "user"
}
] |
Multi-hop Reasoning
|
{
"ground_truth": {
"evidence": [
"Dabnou_sentence_1",
"Tahoua_sentence_1"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 545,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.774",
"value": "?search="
},
{
"operation": "search",
"time": "15.983",
"value": "Dabnou"
},
{
"operation": "Now on",
"time": "18.012",
"value": "Dabnou"
},
{
"operation": "Highlighting",
"time": "42.605",
"value": "Dabnou_sentence_3"
},
{
"operation": "Highlighting",
"time": "50.86",
"value": "Dabnou_sentence_4"
},
{
"operation": "Highlighting",
"time": "57.038",
"value": "Dabnou_sentence_5"
},
{
"operation": "Highlighting",
"time": "66.875",
"value": "Dabnou_sentence_1"
},
{
"operation": "hyperlink",
"time": "68.196",
"value": "Tahoua"
},
{
"operation": "Now on",
"time": "68.803",
"value": "Tahoua"
},
{
"operation": "Highlighting",
"time": "77.297",
"value": "Tahoua_sentence_1"
},
{
"operation": "finish",
"time": "122.495",
"value": "finish"
}
]
}
|
55520
|
The signal tower on Brow Head, which originally used flags and blackballs on masts, was built to give warning of a French invasion during the Napoleonic Wars.
|
{
"evidence": [
"Brow Head_sentence_4",
"Brow Head_sentence_5",
"Napoleonic Wars_sentence_0"
],
"label": "SUPPORT"
}
|
feverous
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nThe signal tower on Brow Head, which originally used flags and blackballs on masts, was built to give warning of a French invasion during the Napoleonic Wars.\n",
"role": "user"
}
] |
Multi-hop Reasoning
|
{
"ground_truth": {
"evidence": [
"Brow Head_sentence_4",
"Brow Head_sentence_5",
"Napoleonic Wars_sentence_0"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 546,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.816",
"value": "?search="
},
{
"operation": "search",
"time": "12.507",
"value": "Brow Head"
},
{
"operation": "Now on",
"time": "13.237",
"value": "Brow Head"
},
{
"operation": "Highlighting",
"time": "25.467",
"value": "Brow Head_sentence_4"
},
{
"operation": "Highlighting",
"time": "44.445",
"value": "Brow Head_sentence_5"
},
{
"operation": "search",
"time": "62.684",
"value": "Napoleonic Wars"
},
{
"operation": "Now on",
"time": "65.389",
"value": "Napoleonic Wars"
},
{
"operation": "Highlighting",
"time": "68.405",
"value": "Napoleonic Wars_sentence_0"
},
{
"operation": "finish",
"time": "90.588",
"value": "finish"
}
]
}
|
88704
|
The Campeones Cup's 2020 edition was canceled due to the COVID-19 pandemic, which started in December 2019.
|
{
"evidence": [
"COVID-19 pandemic_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:\nThe Campeones Cup's 2020 edition was canceled due to the COVID-19 pandemic, which started in December 2019.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"COVID-19 pandemic_sentence_1"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 547,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.907",
"value": "?search="
},
{
"operation": "search",
"time": "155.852",
"value": "Campeones Cup"
},
{
"operation": "Now on",
"time": "157.303",
"value": "Campeones Cup"
},
{
"operation": "Highlighting",
"time": "202.266",
"value": "Campeones Cup_sentence_10"
},
{
"operation": "Highlighting",
"time": "202.267",
"value": "Campeones Cup_title"
},
{
"operation": "Highlighting",
"time": "202.268",
"value": "Campeones Cup_section_1"
},
{
"operation": "hyperlink",
"time": "210.385",
"value": "COVID-19 pandemic"
},
{
"operation": "Now on",
"time": "214.101",
"value": "COVID-19 pandemic"
},
{
"operation": "Highlighting",
"time": "227.978",
"value": "COVID-19 pandemic_sentence_1"
},
{
"operation": "back-button-clicked",
"time": "261.234",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "261.447",
"value": "Campeones Cup"
},
{
"operation": "Highlighting deleted",
"time": "262.938",
"value": "Campeones Cup_sentence_10"
},
{
"operation": "Highlighting deleted",
"time": "262.939",
"value": "Campeones Cup_title"
},
{
"operation": "Highlighting deleted",
"time": "262.941",
"value": "Campeones Cup_section_1"
},
{
"operation": "finish",
"time": "279.65",
"value": "finish"
}
]
}
|
48580
|
Jacques Aubert's (who was born in Paris, the capital and most populous city of France) father Jean was a member of the 24 Violins du Roi until his death, when his son succeeded him, and served within the years until 1746.
|
{
"evidence": [
"Jacques Aubert_sentence_10",
"Jacques Aubert_sentence_12",
"Jacques Aubert_sentence_3",
"Paris_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:\nJacques Aubert's (who was born in Paris, the capital and most populous city of France) father Jean was a member of the 24 Violins du Roi until his death, when his son succeeded him, and served within the years until 1746.\n",
"role": "user"
}
] |
Multi-hop Reasoning
|
{
"ground_truth": {
"evidence": [
"Jacques Aubert_sentence_10",
"Jacques Aubert_sentence_12",
"Jacques Aubert_sentence_3",
"Paris_sentence_2"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 548,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.667",
"value": "?search="
},
{
"operation": "search",
"time": "17.934",
"value": "Jacques Aubert"
},
{
"operation": "Now on",
"time": "19.248",
"value": "Jacques Aubert"
},
{
"operation": "Highlighting",
"time": "33.388",
"value": "Jacques Aubert_sentence_10"
},
{
"operation": "Highlighting",
"time": "57.002",
"value": "Jacques Aubert_sentence_11"
},
{
"operation": "Highlighting deleted",
"time": "74.536",
"value": "Jacques Aubert_sentence_11"
},
{
"operation": "Highlighting",
"time": "81.852",
"value": "Jacques Aubert_sentence_11"
},
{
"operation": "Highlighting",
"time": "84.586",
"value": "Jacques Aubert_sentence_12"
},
{
"operation": "Highlighting deleted",
"time": "118.774",
"value": "Jacques Aubert_sentence_11"
},
{
"operation": "Highlighting",
"time": "133.814",
"value": "Jacques Aubert_sentence_3"
},
{
"operation": "hyperlink",
"time": "139.695",
"value": "Paris"
},
{
"operation": "Now on",
"time": "144.746",
"value": "Paris"
},
{
"operation": "Highlighting",
"time": "148.381",
"value": "Paris_sentence_2"
},
{
"operation": "back-button-clicked",
"time": "149.562",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "149.62",
"value": "Jacques Aubert"
},
{
"operation": "finish",
"time": "170.391",
"value": "finish"
}
]
}
|
exfever_154
|
The AK-47 is a gas-operated assault rifle chambered for the 7.62×39mm cartridge, developed by Vladimir Lenin in the socialist and communist state that spanned much of Eurasia from 1922 to 1991.
|
{
"evidence": [
"AK-47",
"Soviet Union"
],
"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 AK-47 is a gas-operated assault rifle chambered for the 7.62×39mm cartridge, developed by Vladimir Lenin in the socialist and communist state that spanned much of Eurasia from 1922 to 1991.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"AK-47",
"Soviet Union"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 549,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
92845
|
Brinsley is a mining village in England that is significant in the area because of its headstocks which were recetly restored and brought back near its original place.
|
{
"evidence": [
"Brinsley_sentence_8",
"Brinsley_sentence_10",
"Brinsley_sentence_11"
],
"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:\nBrinsley is a mining village in England that is significant in the area because of its headstocks which were recetly restored and brought back near its original place.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Brinsley_sentence_8",
"Brinsley_sentence_10",
"Brinsley_sentence_11"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 550,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.058",
"value": "?search="
},
{
"operation": "search",
"time": "7.889",
"value": "Brinsley"
},
{
"operation": "Now on",
"time": "8.597",
"value": "Brinsley"
},
{
"operation": "Highlighting",
"time": "15.129",
"value": "Brinsley_sentence_2"
},
{
"operation": "Highlighting",
"time": "17.98",
"value": "Brinsley_sentence_8"
},
{
"operation": "Highlighting deleted",
"time": "22.027",
"value": "Brinsley_sentence_2"
},
{
"operation": "Highlighting",
"time": "24.951",
"value": "Brinsley_sentence_10"
},
{
"operation": "Highlighting",
"time": "38.339",
"value": "Brinsley_sentence_11"
},
{
"operation": "finish",
"time": "43.759",
"value": "finish"
}
]
}
|
66150
|
Russian, wild forest reindeer, similar to North American woodland caribou, reached their peak population in the 1800's.
|
{
"evidence": [
"Reindeer in Russia_sentence_40",
"Reindeer in Russia_sentence_5"
],
"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:\nRussian, wild forest reindeer, similar to North American woodland caribou, reached their peak population in the 1800's.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Reindeer in Russia_sentence_40",
"Reindeer in Russia_sentence_5"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 551,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.142",
"value": "?search="
},
{
"operation": "search",
"time": "28.986",
"value": "Contains...Russian wild forest reindeer,"
},
{
"operation": "Now on",
"time": "30.726",
"value": "undefined"
},
{
"operation": "hyperlink",
"time": "46.385",
"value": "Reindeer"
},
{
"operation": "Now on",
"time": "47.588",
"value": "Reindeer"
},
{
"operation": "Page search",
"time": "53.698",
"value": "russian"
},
{
"operation": "hyperlink",
"time": "68.341",
"value": "Russian"
},
{
"operation": "Now on",
"time": "71.915",
"value": "Reindeer"
},
{
"operation": "back-button-clicked",
"time": "111.14",
"value": "back-button-clicked"
},
{
"operation": "back-button-clicked",
"time": "111.156",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "111.875",
"value": "undefined"
},
{
"operation": "search",
"time": "128.109",
"value": "Contains...Russian wild forest reindeer"
},
{
"operation": "Now on",
"time": "129.546",
"value": "undefined"
},
{
"operation": "search",
"time": "166.28",
"value": "Contains...North American woodland caribou"
},
{
"operation": "Now on",
"time": "169.09",
"value": "undefined"
},
{
"operation": "search",
"time": "186.471",
"value": "Contains...Russian, wild forest reindeer, similar to North American woodland caribou, reached their peak population in the 1800's."
},
{
"operation": "Now on",
"time": "188.701",
"value": "undefined"
},
{
"operation": "search",
"time": "207.529",
"value": "Contains...wild forest reindeer"
},
{
"operation": "Now on",
"time": "208.504",
"value": "undefined"
},
{
"operation": "hyperlink",
"time": "220.604",
"value": "Reindeer in Russia"
},
{
"operation": "Now on",
"time": "221.215",
"value": "Reindeer in Russia"
},
{
"operation": "Page search",
"time": "229.927",
"value": "woodland"
},
{
"operation": "Highlighting",
"time": "239.649",
"value": "Reindeer in Russia_sentence_40"
},
{
"operation": "Page search",
"time": "258.664",
"value": "population"
},
{
"operation": "Page search",
"time": "276.548",
"value": "peak population"
},
{
"operation": "Page search",
"time": "288.875",
"value": "population"
},
{
"operation": "Highlighting",
"time": "316.112",
"value": "Reindeer in Russia_sentence_47"
},
{
"operation": "Highlighting deleted",
"time": "320.584",
"value": "Reindeer in Russia_sentence_47"
},
{
"operation": "Highlighting",
"time": "322.04",
"value": "Reindeer in Russia_sentence_52"
},
{
"operation": "Highlighting deleted",
"time": "326.266",
"value": "Reindeer in Russia_sentence_52"
},
{
"operation": "Page search",
"time": "343.546",
"value": "highest population"
},
{
"operation": "Page search",
"time": "361.824",
"value": "population"
},
{
"operation": "Highlighting",
"time": "380.934",
"value": "Reindeer in Russia_sentence_16"
},
{
"operation": "Highlighting",
"time": "380.935",
"value": "Reindeer in Russia_section_1"
},
{
"operation": "Highlighting",
"time": "380.935",
"value": "Reindeer in Russia_section_2"
},
{
"operation": "Highlighting deleted",
"time": "384.18",
"value": "Reindeer in Russia_sentence_16"
},
{
"operation": "Highlighting deleted",
"time": "384.183",
"value": "Reindeer in Russia_section_1"
},
{
"operation": "Highlighting deleted",
"time": "384.185",
"value": "Reindeer in Russia_section_2"
},
{
"operation": "Highlighting",
"time": "391.109",
"value": "Reindeer in Russia_sentence_26"
},
{
"operation": "Highlighting",
"time": "391.11",
"value": "Reindeer in Russia_section_3"
},
{
"operation": "Highlighting deleted",
"time": "403.184",
"value": "Reindeer in Russia_sentence_26"
},
{
"operation": "Highlighting deleted",
"time": "403.185",
"value": "Reindeer in Russia_section_3"
},
{
"operation": "Highlighting",
"time": "409.447",
"value": "Reindeer in Russia_sentence_47"
},
{
"operation": "Highlighting deleted",
"time": "412.796",
"value": "Reindeer in Russia_sentence_47"
},
{
"operation": "Highlighting",
"time": "414.57",
"value": "Reindeer in Russia_sentence_52"
},
{
"operation": "Highlighting deleted",
"time": "428.963",
"value": "Reindeer in Russia_sentence_52"
},
{
"operation": "Highlighting",
"time": "430.422",
"value": "Reindeer in Russia_sentence_53"
},
{
"operation": "Highlighting",
"time": "430.423",
"value": "Reindeer in Russia_section_6"
},
{
"operation": "Highlighting deleted",
"time": "435.945",
"value": "Reindeer in Russia_sentence_53"
},
{
"operation": "Highlighting deleted",
"time": "435.946",
"value": "Reindeer in Russia_section_6"
},
{
"operation": "Page search",
"time": "456.712",
"value": "1800"
},
{
"operation": "Page search",
"time": "463.533",
"value": "18"
},
{
"operation": "Highlighting",
"time": "469.727",
"value": "Reindeer in Russia_sentence_5"
},
{
"operation": "Highlighting",
"time": "485.729",
"value": "Reindeer in Russia_sentence_18"
},
{
"operation": "Highlighting",
"time": "485.73",
"value": "Reindeer in Russia_section_1"
},
{
"operation": "Highlighting",
"time": "485.73",
"value": "Reindeer in Russia_section_2"
},
{
"operation": "Highlighting deleted",
"time": "500.2",
"value": "Reindeer in Russia_sentence_18"
},
{
"operation": "Highlighting deleted",
"time": "500.202",
"value": "Reindeer in Russia_section_1"
},
{
"operation": "Highlighting deleted",
"time": "500.203",
"value": "Reindeer in Russia_section_2"
},
{
"operation": "Highlighting",
"time": "508.138",
"value": "Reindeer in Russia_sentence_67"
},
{
"operation": "Highlighting",
"time": "508.139",
"value": "Reindeer in Russia_section_8"
},
{
"operation": "Highlighting deleted",
"time": "513.184",
"value": "Reindeer in Russia_sentence_67"
},
{
"operation": "Highlighting deleted",
"time": "513.185",
"value": "Reindeer in Russia_section_8"
},
{
"operation": "finish",
"time": "527.808",
"value": "finish"
}
]
}
|
65277
|
Mills, born in Kensett in White County, attended public schools in Kensett but graduated as valedictorian from Searcy High School in the county seat of White County.
|
{
"evidence": [
"Wilbur Mills_sentence_9",
"Wilbur Mills_sentence_11"
],
"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:\nMills, born in Kensett in White County, attended public schools in Kensett but graduated as valedictorian from Searcy High School in the county seat of White County.\n",
"role": "user"
}
] |
Search terms not in claim
|
{
"ground_truth": {
"evidence": [
"Wilbur Mills_sentence_9",
"Wilbur Mills_sentence_11"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 552,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.595",
"value": "undefined"
},
{
"operation": "search",
"time": "78.28",
"value": "Wilbur Mills"
},
{
"operation": "Now on",
"time": "79.867",
"value": "Wilbur Mills"
},
{
"operation": "Highlighting",
"time": "97.15",
"value": "Wilbur Mills_sentence_9"
},
{
"operation": "Highlighting",
"time": "107.358",
"value": "Wilbur Mills_sentence_10"
},
{
"operation": "Highlighting deleted",
"time": "109.885",
"value": "Wilbur Mills_sentence_10"
},
{
"operation": "Highlighting",
"time": "111.044",
"value": "Wilbur Mills_sentence_11"
},
{
"operation": "finish",
"time": "126.173",
"value": "finish"
}
]
}
|
43427
|
Rahah railway station in Iran includes Eslamshahrtoward Tehran and Shuttle bus.
|
{
"evidence": [
"Eslamshahr railway station_sentence_0",
"Islamic Republic of Iran Railways_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:\nRahah railway station in Iran includes Eslamshahrtoward Tehran and Shuttle bus.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Eslamshahr railway station_sentence_0",
"Islamic Republic of Iran Railways_sentence_0"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 553,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.15",
"value": "?search="
},
{
"operation": "search",
"time": "10.038",
"value": "Contains...Rahah railway station"
},
{
"operation": "Now on",
"time": "10.757",
"value": "undefined"
},
{
"operation": "search",
"time": "59.563",
"value": "Islamic Republic of Iran Railways"
},
{
"operation": "Now on",
"time": "61.363",
"value": "Islamic Republic of Iran Railways"
},
{
"operation": "Page search",
"time": "68.693",
"value": "rahah"
},
{
"operation": "Page search",
"time": "93.646",
"value": "eslamshah"
},
{
"operation": "Page search",
"time": "147.221",
"value": "eslam"
},
{
"operation": "search",
"time": "196.296",
"value": "Transport in Iran"
},
{
"operation": "Now on",
"time": "197.526",
"value": "Transport in Iran"
},
{
"operation": "Page search",
"time": "207.677",
"value": "eslamshah"
},
{
"operation": "search",
"time": "231.014",
"value": "Eslamshahr railway station"
},
{
"operation": "Now on",
"time": "231.921",
"value": "Eslamshahr railway station"
},
{
"operation": "Highlighting",
"time": "238.269",
"value": "Eslamshahr railway station_sentence_0"
},
{
"operation": "hyperlink",
"time": "273.181",
"value": "IRI Railway"
},
{
"operation": "Now on",
"time": "274.622",
"value": "Islamic Republic of Iran Railways"
},
{
"operation": "Highlighting",
"time": "280.9",
"value": "Islamic Republic of Iran Railways_sentence_0"
},
{
"operation": "finish",
"time": "286.255",
"value": "finish"
}
]
}
|
exfever_176
|
Machine Gun Kelly is an American dancer whose his third studio album was released on May 12, 2017, by Bad Boy, Interscope Records, and EST 19XX.
|
{
"evidence": [
"Machine Gun Kelly (rapper)",
"Bloom (Machine Gun Kelly album)"
],
"label": "REFUTE"
}
|
exfever
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nMachine Gun Kelly is an American dancer whose his third studio album was released on May 12, 2017, by Bad Boy, Interscope Records, and EST 19XX.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Machine Gun Kelly (rapper)",
"Bloom (Machine Gun Kelly album)"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 554,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
3043
|
To the Nameless Dead is the fifth studio album by the American white metal band Primordial, released in 2009 and it is also available in a limited edition as a digibook.
|
{
"evidence": [
"To the Nameless Dead_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:\nTo the Nameless Dead is the fifth studio album by the American white metal band Primordial, released in 2009 and it is also available in a limited edition as a digibook.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"To the Nameless Dead_sentence_0"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 555,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "search",
"time": "12.037",
"value": "\"to the nameless dead\""
},
{
"operation": "Highlighting",
"time": "22.576",
"value": "To the Nameless Dead_sentence_0"
},
{
"operation": "hyperlink",
"time": "44.717",
"value": "To The Nameless Dead"
},
{
"operation": "finish",
"time": "52.215",
"value": "finish"
}
]
}
|
exfever_496
|
Lyme disease's symptoms may include a disorder that affects the joints, with symptoms generally including stiffness.
|
{
"evidence": [
"Lyme disease",
"Arthritis"
],
"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:\nLyme disease's symptoms may include a disorder that affects the joints, with symptoms generally including stiffness.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Lyme disease",
"Arthritis"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 556,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
73787
|
In the 2015–16 Campionato Sammarinese di Calcio (which was the thirty first season since its establishment), the top three teams from each group moved on to a play-off, which was played in a double elimination format with both group winners earning byes in the first and second round, and determined the season's champion.
|
{
"evidence": [
"2015–16 Campionato Sammarinese di Calcio_sentence_0",
"2015–16 Campionato Sammarinese di Calcio_sentence_8",
"2015–16 Campionato Sammarinese di Calcio_sentence_9"
],
"label": "SUPPORT"
}
|
feverous
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nIn the 2015–16 Campionato Sammarinese di Calcio (which was the thirty first season since its establishment), the top three teams from each group moved on to a play-off, which was played in a double elimination format with both group winners earning byes in the first and second round, and determined the season's champion.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"2015–16 Campionato Sammarinese di Calcio_sentence_0",
"2015–16 Campionato Sammarinese di Calcio_sentence_8",
"2015–16 Campionato Sammarinese di Calcio_sentence_9"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 557,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.894",
"value": "?search="
},
{
"operation": "search",
"time": "676.498",
"value": "2015-16 Campionato Sammarinese di Calcio"
},
{
"operation": "Now on",
"time": "677.713",
"value": "2015–16 Campionato Sammarinese di Calcio"
},
{
"operation": "Highlighting",
"time": "693.906",
"value": "2015–16 Campionato Sammarinese di Calcio_sentence_0"
},
{
"operation": "Highlighting",
"time": "706.51",
"value": "2015–16 Campionato Sammarinese di Calcio_sentence_8"
},
{
"operation": "Highlighting",
"time": "716.338",
"value": "2015–16 Campionato Sammarinese di Calcio_sentence_9"
},
{
"operation": "finish",
"time": "767.835",
"value": "finish"
}
]
}
|
exfever_426
|
Kazakhstan is a transcontinental country located mainly in Central Asia that borders the world's most populous country in South Asia.
|
{
"evidence": [
"Kazakhstan",
"China"
],
"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:\nKazakhstan is a transcontinental country located mainly in Central Asia that borders the world's most populous country in South Asia.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Kazakhstan",
"China"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 558,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
79314
|
Seniors with a life insurance policy are eligible for Medicaid and Orestis created a model for middle-class people to pay for long-term care when they cannot pay out of pocket or through Medicaid.
|
{
"evidence": [
"Medicaid_sentence_210"
],
"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:\nSeniors with a life insurance policy are eligible for Medicaid and Orestis created a model for middle-class people to pay for long-term care when they cannot pay out of pocket or through Medicaid.\n",
"role": "user"
}
] |
Search terms not in claim
|
{
"ground_truth": {
"evidence": [
"Medicaid_sentence_210"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 559,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.97",
"value": "?search="
},
{
"operation": "search",
"time": "14.982",
"value": "Medicaid and Orestis"
},
{
"operation": "Now on",
"time": "16.116",
"value": "undefined"
},
{
"operation": "hyperlink",
"time": "22.256",
"value": "medicare and orestes"
},
{
"operation": "Now on",
"time": "22.963",
"value": "undefined"
},
{
"operation": "hyperlink",
"time": "26.298",
"value": "medicine and orestes"
},
{
"operation": "Now on",
"time": "27.746",
"value": "undefined"
},
{
"operation": "search",
"time": "32.977",
"value": "medicaid"
},
{
"operation": "Now on",
"time": "36.222",
"value": "Medicaid"
},
{
"operation": "Page search",
"time": "156.785",
"value": "life insurance"
},
{
"operation": "Page search",
"time": "158.479",
"value": "life insurance"
},
{
"operation": "search",
"time": "232.686",
"value": "orestis"
},
{
"operation": "Now on",
"time": "235.02",
"value": "Orestis"
},
{
"operation": "back-button-clicked",
"time": "269.665",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "269.81",
"value": "Medicaid"
},
{
"operation": "hyperlink",
"time": "273.278",
"value": "Medicare (United States)"
},
{
"operation": "Now on",
"time": "276.012",
"value": "Medicare (United States)"
},
{
"operation": "Page search",
"time": "295.448",
"value": "middle-class people"
},
{
"operation": "Page search",
"time": "317.747",
"value": "life insurance"
},
{
"operation": "search",
"time": "358.17",
"value": "life insurance policy"
},
{
"operation": "Now on",
"time": "363.026",
"value": "Life insurance"
},
{
"operation": "Page search",
"time": "373.379",
"value": "medicaid"
},
{
"operation": "search",
"time": "426.176",
"value": "Medicaid"
},
{
"operation": "Now on",
"time": "427.858",
"value": "Medicaid"
},
{
"operation": "Page search",
"time": "610.296",
"value": "long-term care"
},
{
"operation": "Highlighting",
"time": "676.836",
"value": "Medicaid_sentence_210"
},
{
"operation": "finish",
"time": "689.53",
"value": "finish"
}
]
}
|
exfever_3872
|
Joel Daniel Coen and Ethan Jesse Coen are American filmmakers whose most acclaimed works include the 1990 American neo-noir gangster film starring Gabriel Byrne, Marcia Gay Harden, John Turturro, Jon Polito, J. E. Freeman, and Albert Finney.
|
{
"evidence": [
"Coen brothers",
"Miller's Crossing"
],
"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:\nJoel Daniel Coen and Ethan Jesse Coen are American filmmakers whose most acclaimed works include the 1990 American neo-noir gangster film starring Gabriel Byrne, Marcia Gay Harden, John Turturro, Jon Polito, J. E. Freeman, and Albert Finney.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Coen brothers",
"Miller's Crossing"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 560,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
29900
|
Devin Oosthuizen (born 28 May 1988) started both legs of their promotion play-off matches against the Pumas, and his most prolific try-scoring campaign was the 2011 Vodacom Cup
|
{
"evidence": [
"Devin Oosthuizen_sentence_0",
"Devin Oosthuizen_sentence_15",
"Devin Oosthuizen_sentence_17"
],
"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:\nDevin Oosthuizen (born 28 May 1988) started both legs of their promotion play-off matches against the Pumas, and his most prolific try-scoring campaign was the 2011 Vodacom Cup\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Devin Oosthuizen_sentence_0",
"Devin Oosthuizen_sentence_15",
"Devin Oosthuizen_sentence_17"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 561,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.059",
"value": "?search="
},
{
"operation": "search",
"time": "3.381",
"value": "Devin Oosthuizen"
},
{
"operation": "Now on",
"time": "4.243",
"value": "Devin Oosthuizen"
},
{
"operation": "Highlighting",
"time": "5.51",
"value": "Devin Oosthuizen_sentence_0"
},
{
"operation": "Highlighting",
"time": "13.18",
"value": "Devin Oosthuizen_sentence_7"
},
{
"operation": "Highlighting",
"time": "13.182",
"value": "Devin Oosthuizen_section_2"
},
{
"operation": "Highlighting deleted",
"time": "15.295",
"value": "Devin Oosthuizen_sentence_7"
},
{
"operation": "Highlighting deleted",
"time": "15.299",
"value": "Devin Oosthuizen_section_2"
},
{
"operation": "Highlighting",
"time": "20.226",
"value": "Devin Oosthuizen_sentence_15"
},
{
"operation": "Highlighting",
"time": "29.134",
"value": "Devin Oosthuizen_sentence_17"
},
{
"operation": "finish",
"time": "35.989",
"value": "finish"
}
]
}
|
79827
|
Chill point the downtown district is original market of the Stop & Shop supermarket chain was on the corner of Northern Artery and McGrath Highway but in 2008, it would be dismantled and sold to Daewoo Shipbuilding & Marine Engineering of South Korea and relocated to Mangalia, Romania.
|
{
"evidence": [
"Quincy Point_sentence_18"
],
"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:\nChill point the downtown district is original market of the Stop & Shop supermarket chain was on the corner of Northern Artery and McGrath Highway but in 2008, it would be dismantled and sold to Daewoo Shipbuilding & Marine Engineering of South Korea and relocated to Mangalia, Romania.\n",
"role": "user"
}
] |
Search terms not in claim
|
{
"ground_truth": {
"evidence": [
"Quincy Point_sentence_18"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 562,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "2.351",
"value": "?search="
},
{
"operation": "search",
"time": "19.979",
"value": "Contains...Chill point"
},
{
"operation": "Now on",
"time": "21.7",
"value": "undefined"
},
{
"operation": "search",
"time": "218.274",
"value": "Quincy Point"
},
{
"operation": "Now on",
"time": "219.115",
"value": "Quincy Point"
},
{
"operation": "Page search",
"time": "226.592",
"value": "stop"
},
{
"operation": "Highlighting",
"time": "230.404",
"value": "Quincy Point_sentence_18"
},
{
"operation": "Page search",
"time": "251.119",
"value": "mangalia"
},
{
"operation": "finish",
"time": "275.23",
"value": "finish"
}
]
}
|
13050
|
David Bar Katz is an British screenwriter, playwright, author, director and philanthropist who is a member of Israel Horovitz's New York Playwright's Lab and New York City's LAByrinth Theater Company.
|
{
"evidence": [
"David Bar Katz_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:\nDavid Bar Katz is an British screenwriter, playwright, author, director and philanthropist who is a member of Israel Horovitz's New York Playwright's Lab and New York City's LAByrinth Theater Company.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"David Bar Katz_sentence_0"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 563,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.202",
"value": "?search="
},
{
"operation": "search",
"time": "11.927",
"value": "David Bar Katz"
},
{
"operation": "Now on",
"time": "13.216",
"value": "David Bar Katz"
},
{
"operation": "Highlighting",
"time": "16.064",
"value": "David Bar Katz_sentence_0"
},
{
"operation": "finish",
"time": "28.579",
"value": "finish"
}
]
}
|
65109
|
ICAM3 is an intercellular adhesion molecule that functions as both an adhesion and a potent signaling molecule.
|
{
"evidence": [
"ICAM3_sentence_0",
"ICAM3_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:\nICAM3 is an intercellular adhesion molecule that functions as both an adhesion and a potent signaling molecule.\n",
"role": "user"
}
] |
Multi-hop Reasoning
|
{
"ground_truth": {
"evidence": [
"ICAM3_sentence_0",
"ICAM3_sentence_4"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 564,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.975",
"value": "?search="
},
{
"operation": "search",
"time": "44.003",
"value": "ICAM3"
},
{
"operation": "Now on",
"time": "45.915",
"value": "ICAM3"
},
{
"operation": "Highlighting",
"time": "62.038",
"value": "ICAM3_sentence_0"
},
{
"operation": "Highlighting",
"time": "97.951",
"value": "ICAM3_sentence_1"
},
{
"operation": "Highlighting deleted",
"time": "108.772",
"value": "ICAM3_sentence_1"
},
{
"operation": "Page search",
"time": "125.473",
"value": "adhesion"
},
{
"operation": "Page search",
"time": "142.912",
"value": "potent signaling molecule"
},
{
"operation": "Highlighting",
"time": "169.092",
"value": "ICAM3_sentence_4"
},
{
"operation": "finish",
"time": "180.975",
"value": "finish"
}
]
}
|
exfever_6884
|
Broadcast News is a 1987 American romantic comedy-drama film starring an American actress and former fashion model known for her role as Holly Goodhead in the James Bond film Moonraker.
|
{
"evidence": [
"Broadcast News (film)",
"Lois Chiles"
],
"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:\nBroadcast News is a 1987 American romantic comedy-drama film starring an American actress and former fashion model known for her role as Holly Goodhead in the James Bond film Moonraker.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Broadcast News (film)",
"Lois Chiles"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 565,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_5125
|
The Royal Tenenbaums is a 2005 American comedy-drama film directed by an American filmmaker.
|
{
"evidence": [
"The Royal Tenenbaums",
"Wes Anderson"
],
"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 Royal Tenenbaums is a 2005 American comedy-drama film directed by an American filmmaker.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"The Royal Tenenbaums",
"Wes Anderson"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 566,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
5765
|
Erick Harris signed with the Arena Football League's Manchester Wolves and recorded three interceptions before moving to the Cleveland Gladiators.
|
{
"evidence": [
"Erick Harris_sentence_17",
"Erick Harris_sentence_18",
"Erick Harris_sentence_19"
],
"label": "SUPPORT"
}
|
feverous
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nErick Harris signed with the Arena Football League's Manchester Wolves and recorded three interceptions before moving to the Cleveland Gladiators.\n",
"role": "user"
}
] |
Entity Disambiguation
|
{
"ground_truth": {
"evidence": [
"Erick Harris_sentence_17",
"Erick Harris_sentence_18",
"Erick Harris_sentence_19"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 567,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.736",
"value": "?search="
},
{
"operation": "search",
"time": "9.187",
"value": "Erik Harris"
},
{
"operation": "Now on",
"time": "10.193",
"value": "Erik Harris"
},
{
"operation": "Page search",
"time": "38.513",
"value": "Arena Football League"
},
{
"operation": "Page search",
"time": "50.52",
"value": "Machester Wolves"
},
{
"operation": "Page search",
"time": "54.864",
"value": "Manchester Wolves"
},
{
"operation": "search",
"time": "78.328",
"value": "Erick Harris"
},
{
"operation": "Now on",
"time": "78.841",
"value": "Erick Harris"
},
{
"operation": "Highlighting",
"time": "106.192",
"value": "Erick Harris_sentence_17"
},
{
"operation": "Highlighting",
"time": "113.991",
"value": "Erick Harris_sentence_18"
},
{
"operation": "Highlighting",
"time": "124.054",
"value": "Erick Harris_sentence_19"
},
{
"operation": "finish",
"time": "140.358",
"value": "finish"
}
]
}
|
49443
|
Musicane offers commissions for music they help sell, but they are not an MP3 retail partner with The Orchard
|
{
"evidence": [
"Musicane_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:\nMusicane offers commissions for music they help sell, but they are not an MP3 retail partner with The Orchard\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Musicane_sentence_5"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 568,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "4.951",
"value": "?search="
},
{
"operation": "search",
"time": "11.484",
"value": "Musicane"
},
{
"operation": "Now on",
"time": "12.71",
"value": "Musicane"
},
{
"operation": "Highlighting",
"time": "17.413",
"value": "Musicane_sentence_5"
},
{
"operation": "finish",
"time": "35.651",
"value": "finish"
}
]
}
|
93106
|
Their first single, "That's What Boys Do", was a hit and was also used in The Powerpuff Girls Series, and later that year, was used in The Hot Chick and in 2002 after much success with the group, they toured alongside Aaron Carter on his "Oh Aaron tour" as a part of a Jive Records collaboration.
|
{
"evidence": [
"Nancarrow_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:\nTheir first single, \"That's What Boys Do\", was a hit and was also used in The Powerpuff Girls Series, and later that year, was used in The Hot Chick and in 2002 after much success with the group, they toured alongside Aaron Carter on his \"Oh Aaron tour\" as a part of a Jive Records collaboration.\n",
"role": "user"
}
] |
Search terms not in claim
|
{
"ground_truth": {
"evidence": [
"Nancarrow_sentence_0"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 569,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.783",
"value": "?search="
},
{
"operation": "search",
"time": "18.921",
"value": "Contains...That's What Boys Do"
},
{
"operation": "Now on",
"time": "24.004",
"value": "undefined"
},
{
"operation": "search",
"time": "119.932",
"value": "Nancarrow"
},
{
"operation": "Now on",
"time": "121.432",
"value": "Nancarrow"
},
{
"operation": "search",
"time": "185.613",
"value": "Nancarrow"
},
{
"operation": "Now on",
"time": "186.322",
"value": "Nancarrow"
},
{
"operation": "Highlighting",
"time": "204.852",
"value": "Nancarrow_sentence_0"
},
{
"operation": "back-button-clicked",
"time": "229.43",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "229.498",
"value": "Nancarrow"
},
{
"operation": "Now on",
"time": "241.779",
"value": "Nancarrow"
},
{
"operation": "finish",
"time": "248.033",
"value": "finish"
}
]
}
|
exfever_6904
|
Leslie Howard was an actor who played Ashley Wilkes in a 1939 American epic historical romance film.
|
{
"evidence": [
"Leslie Howard (actor)",
"Gone with the Wind (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:\nLeslie Howard was an actor who played Ashley Wilkes in a 1939 American epic historical romance film.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Leslie Howard (actor)",
"Gone with the Wind (film)"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 570,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
43611
|
The story in The Cure for Everything, by writer Severna Park, includes two characters named Maria and N'Lykli, as well as a woman from a North American tribe who could be the source of a hundred new vaccines.
|
{
"evidence": [
"The Cure for Everything_sentence_0",
"The Cure for Everything_sentence_3",
"The Cure for Everything_sentence_5",
"The Cure for Everything_sentence_8"
],
"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 story in The Cure for Everything, by writer Severna Park, includes two characters named Maria and N'Lykli, as well as a woman from a North American tribe who could be the source of a hundred new vaccines.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"The Cure for Everything_sentence_0",
"The Cure for Everything_sentence_3",
"The Cure for Everything_sentence_5",
"The Cure for Everything_sentence_8"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 571,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.631",
"value": "?search="
},
{
"operation": "search",
"time": "180.805",
"value": "The Cure for Everything"
},
{
"operation": "Now on",
"time": "181.698",
"value": "The Cure for Everything"
},
{
"operation": "Highlighting",
"time": "392.749",
"value": "The Cure for Everything_sentence_0"
},
{
"operation": "Highlighting",
"time": "408.709",
"value": "The Cure for Everything_sentence_3"
},
{
"operation": "Highlighting",
"time": "417.831",
"value": "The Cure for Everything_sentence_5"
},
{
"operation": "Highlighting",
"time": "443.829",
"value": "The Cure for Everything_sentence_8"
},
{
"operation": "Page search",
"time": "461.361",
"value": "tribe"
},
{
"operation": "search",
"time": "487.772",
"value": "Xingu"
},
{
"operation": "Now on",
"time": "488.729",
"value": "Xingu"
},
{
"operation": "back-button-clicked",
"time": "494.117",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "494.141",
"value": "The Cure for Everything"
},
{
"operation": "Page search",
"time": "510.137",
"value": "america"
},
{
"operation": "finish",
"time": "514.943",
"value": "finish"
}
]
}
|
exfever_7304
|
Virginia, in the Mid-Atlantic and Southeastern regions of the United States, is a constituent political entity, of which there are currently 60.
|
{
"evidence": [
"Virginia",
"U.S. state"
],
"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:\nVirginia, in the Mid-Atlantic and Southeastern regions of the United States, is a constituent political entity, of which there are currently 60.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Virginia",
"U.S. state"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 572,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
56868
|
Maldroid became popular on local San Francisco modern rock (format consists primarily of the alternative rock genre) station and earned them an an appearance on Good Morning America
|
{
"evidence": [
"Maldroid_sentence_11",
"Maldroid_sentence_12",
"Modern rock_sentence_7"
],
"label": "SUPPORT"
}
|
feverous
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nMaldroid became popular on local San Francisco modern rock (format consists primarily of the alternative rock genre) station and earned them an an appearance on Good Morning America\n",
"role": "user"
}
] |
Multi-hop Reasoning
|
{
"ground_truth": {
"evidence": [
"Maldroid_sentence_11",
"Maldroid_sentence_12",
"Modern rock_sentence_7"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 573,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "4.704",
"value": "?search="
},
{
"operation": "search",
"time": "15.113",
"value": "Contains...maldroid"
},
{
"operation": "Now on",
"time": "15.857",
"value": "undefined"
},
{
"operation": "hyperlink",
"time": "18.007",
"value": "Maldroid"
},
{
"operation": "Now on",
"time": "18.914",
"value": "Maldroid"
},
{
"operation": "Highlighting",
"time": "43.852",
"value": "Maldroid_sentence_11"
},
{
"operation": "Highlighting",
"time": "46.265",
"value": "Maldroid_sentence_12"
},
{
"operation": "hyperlink",
"time": "51.336",
"value": "modern rock"
},
{
"operation": "Now on",
"time": "52.127",
"value": "Modern rock"
},
{
"operation": "Highlighting",
"time": "63.947",
"value": "Modern rock_sentence_7"
},
{
"operation": "finish",
"time": "72.085",
"value": "finish"
}
]
}
|
exfever_6583
|
Patrick Galen Dempsey is an African actor who was successful in landing a lead role in a 2002 American romantic comedy film directed by Andy Tennant and starring Reese Witherspoon.
|
{
"evidence": [
"Patrick Dempsey",
"Sweet Home Alabama (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:\nPatrick Galen Dempsey is an African actor who was successful in landing a lead role in a 2002 American romantic comedy film directed by Andy Tennant and starring Reese Witherspoon.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Patrick Dempsey",
"Sweet Home Alabama (film)"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 574,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_7904
|
Finding Dory has earned $488.3 million in the United States and Canada.
|
{
"evidence": [
"79th Academy Awards",
"Ellen DeGeneres",
"Pixar"
],
"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:\nFinding Dory has earned $488.3 million in the United States and Canada.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"79th Academy Awards",
"Ellen DeGeneres",
"Pixar"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 575,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
56678
|
As a part of this campaign, Michael B Simon managed stockpiles of data on voter behavior and preferences and he has been a guest speaker at Harvard's Kennedy School of Government and the World Economic Forum of Young Global Leaders.
|
{
"evidence": [
"Michael B Simon_sentence_5",
"Michael B Simon_sentence_8"
],
"label": "SUPPORT"
}
|
feverous
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nAs a part of this campaign, Michael B Simon managed stockpiles of data on voter behavior and preferences and he has been a guest speaker at Harvard's Kennedy School of Government and the World Economic Forum of Young Global Leaders.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Michael B Simon_sentence_5",
"Michael B Simon_sentence_8"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 576,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.173",
"value": "?search="
},
{
"operation": "search",
"time": "11.393",
"value": "Michael B Simon"
},
{
"operation": "Now on",
"time": "11.964",
"value": "Michael B Simon"
},
{
"operation": "Page search",
"time": "26.516",
"value": "stockp"
},
{
"operation": "Highlighting",
"time": "29.678",
"value": "Michael B Simon_sentence_5"
},
{
"operation": "Page search",
"time": "38.132",
"value": "speak"
},
{
"operation": "Highlighting",
"time": "47.665",
"value": "Michael B Simon_sentence_8"
},
{
"operation": "finish",
"time": "53.458",
"value": "finish"
}
]
}
|
29281
|
During 2015 North Tyneside Metropolitan Borough Council election held on 7 May there were a total of 72 candidates, the Labour Party fielded 20 candidates and 1 candidate represented National Front.
|
{
"evidence": [
"2015 North Tyneside Metropolitan Borough Council election_sentence_0",
"2015 North Tyneside Metropolitan Borough Council election_sentence_5",
"2015 North Tyneside Metropolitan Borough Council election_sentence_6",
"2015 North Tyneside Metropolitan Borough Council election_sentence_8"
],
"label": "SUPPORT"
}
|
feverous
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nDuring 2015 North Tyneside Metropolitan Borough Council election held on 7 May there were a total of 72 candidates, the Labour Party fielded 20 candidates and 1 candidate represented National Front.\n",
"role": "user"
}
] |
Entity Disambiguation
|
{
"ground_truth": {
"evidence": [
"2015 North Tyneside Metropolitan Borough Council election_sentence_0",
"2015 North Tyneside Metropolitan Borough Council election_sentence_5",
"2015 North Tyneside Metropolitan Borough Council election_sentence_6",
"2015 North Tyneside Metropolitan Borough Council election_sentence_8"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 577,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.505",
"value": "undefined"
},
{
"operation": "search",
"time": "15.34",
"value": "2015 North Tyneside Metropolitan Borough Council election"
},
{
"operation": "Now on",
"time": "16.818",
"value": "2015 North Tyneside Metropolitan Borough Council election"
},
{
"operation": "Highlighting",
"time": "19.296",
"value": "2015 North Tyneside Metropolitan Borough Council election_sentence_0"
},
{
"operation": "Highlighting",
"time": "31.307",
"value": "2015 North Tyneside Metropolitan Borough Council election_sentence_5"
},
{
"operation": "Highlighting",
"time": "31.937",
"value": "2015 North Tyneside Metropolitan Borough Council election_sentence_6"
},
{
"operation": "Highlighting",
"time": "34.365",
"value": "2015 North Tyneside Metropolitan Borough Council election_sentence_8"
},
{
"operation": "finish",
"time": "74.946",
"value": "finish"
}
]
}
|
26301
|
Alan Cockeran (September 7, 1894 –July 14, 1961, born in Scotland) emigrated to Canada in 1933 and he found employment with Toronto-Dominion Bank in Toronto, then later became the Prime Minister in 1945.
|
{
"evidence": [
"Alan Cockeram_sentence_0",
"Alan Cockeram_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:\nAlan Cockeran (September 7, 1894 –July 14, 1961, born in Scotland) emigrated to Canada in 1933 and he found employment with Toronto-Dominion Bank in Toronto, then later became the Prime Minister in 1945.\n",
"role": "user"
}
] |
Search terms not in claim
|
{
"ground_truth": {
"evidence": [
"Alan Cockeram_sentence_0",
"Alan Cockeram_sentence_2"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 578,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.725",
"value": "?search="
},
{
"operation": "search",
"time": "14.544",
"value": "Alan Cockeran"
},
{
"operation": "Now on",
"time": "15.788",
"value": "undefined"
},
{
"operation": "hyperlink",
"time": "19.783",
"value": "Alan Cockeram"
},
{
"operation": "Now on",
"time": "20.2",
"value": "Alan Cockeram"
},
{
"operation": "back-button-clicked",
"time": "27.568",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "28.598",
"value": "undefined"
},
{
"operation": "hyperlink",
"time": "46.14",
"value": "Alan Cockeram"
},
{
"operation": "Now on",
"time": "46.423",
"value": "Alan Cockeram"
},
{
"operation": "Highlighting",
"time": "53.779",
"value": "Alan Cockeram_sentence_2"
},
{
"operation": "Highlighting",
"time": "97.458",
"value": "Alan Cockeram_sentence_11"
},
{
"operation": "Highlighting",
"time": "99.608",
"value": "Alan Cockeram_sentence_0"
},
{
"operation": "Highlighting deleted",
"time": "104.7",
"value": "Alan Cockeram_sentence_2"
},
{
"operation": "Highlighting deleted",
"time": "106.683",
"value": "Alan Cockeram_sentence_11"
},
{
"operation": "Highlighting",
"time": "149.937",
"value": "Alan Cockeram_sentence_2"
},
{
"operation": "finish",
"time": "179.696",
"value": "finish"
}
]
}
|
exfever_3867
|
Emilio Estévez, the son of Martin Sheen, is an American actor known for a 1984 American science fiction black comedy film starring Harry Dean Stanton and Emilio Estevez with a supporting cast, an American actress who debuted in The Great Silence.
|
{
"evidence": [
"Emilio Estevez",
"Repo Man (film)",
"Vonetta McGee"
],
"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:\nEmilio Estévez, the son of Martin Sheen, is an American actor known for a 1984 American science fiction black comedy film starring Harry Dean Stanton and Emilio Estevez with a supporting cast, an American actress who debuted in The Great Silence.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Emilio Estevez",
"Repo Man (film)",
"Vonetta McGee"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 579,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
74787
|
Thomas Soar, born 3 September 1845, in Whitemoor (which is an area in the City of Nottingham, England), had a batting average of 13.38.
|
{
"evidence": [
"Thomas Soar_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:\nThomas Soar, born 3 September 1845, in Whitemoor (which is an area in the City of Nottingham, England), had a batting average of 13.38.\n",
"role": "user"
}
] |
Numerical Reasoning
|
{
"ground_truth": {
"evidence": [
"Thomas Soar_sentence_0"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 580,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.983",
"value": "?search="
},
{
"operation": "search",
"time": "11.335",
"value": "Thomas Soar"
},
{
"operation": "Now on",
"time": "12.389",
"value": "Thomas Soar"
},
{
"operation": "Highlighting",
"time": "18.699",
"value": "Thomas Soar_sentence_0"
},
{
"operation": "finish",
"time": "31.512",
"value": "finish"
}
]
}
|
25649
|
Johann Adolf Köster (c. 1550 – 1630) was a pastor, teacher and historian from Bamberg, Germany who chronicled the medieval history of Dithmarschen.
|
{
"evidence": [
"Neocorus_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:\nJohann Adolf Köster (c. 1550 – 1630) was a pastor, teacher and historian from Bamberg, Germany who chronicled the medieval history of Dithmarschen.\n",
"role": "user"
}
] |
Search terms not in claim
|
{
"ground_truth": {
"evidence": [
"Neocorus_sentence_1"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 581,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.899",
"value": "?search="
},
{
"operation": "search",
"time": "5.282",
"value": "Contains...Johann Adolf Köster"
},
{
"operation": "Now on",
"time": "6.704",
"value": "undefined"
},
{
"operation": "hyperlink",
"time": "8.692",
"value": "Köster"
},
{
"operation": "Now on",
"time": "9.192",
"value": "Köster"
},
{
"operation": "hyperlink",
"time": "11.037",
"value": "Adolf Köster"
},
{
"operation": "Now on",
"time": "11.547",
"value": "Adolf Köster"
},
{
"operation": "Page search",
"time": "46.868",
"value": "Dithmarschen"
},
{
"operation": "Page search",
"time": "47.324",
"value": "Dithmarschen"
},
{
"operation": "search",
"time": "49.196",
"value": "Contains...Dithmarschen"
},
{
"operation": "Now on",
"time": "49.802",
"value": "undefined"
},
{
"operation": "search",
"time": "81.825",
"value": "Neocorus"
},
{
"operation": "Now on",
"time": "82.382",
"value": "Neocorus"
},
{
"operation": "Highlighting",
"time": "84.416",
"value": "Neocorus_sentence_1"
},
{
"operation": "Page search",
"time": "96.665",
"value": "bamb"
},
{
"operation": "Page search",
"time": "96.899",
"value": "bamb"
},
{
"operation": "finish",
"time": "106.586",
"value": "finish"
}
]
}
|
exfever_8576
|
Kimila Ann Basinger was the wealthiest American actress in America.
|
{
"evidence": [
"Kim Basinger",
"Never Say Never Again"
],
"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:\nKimila Ann Basinger was the wealthiest American actress in America.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Kim Basinger",
"Never Say Never Again"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 582,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
20301
|
Fonogram 2007 winner Magdi Ruzsa, outvoting ten other artists, represented Hungary in the Eurovision Song Contest 2007.
|
{
"evidence": [
"Magdi Rúzsa_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:\nFonogram 2007 winner Magdi Ruzsa, outvoting ten other artists, represented Hungary in the Eurovision Song Contest 2007.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Magdi Rúzsa_sentence_3"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 583,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.021",
"value": "?search="
},
{
"operation": "search",
"time": "23.258",
"value": "Magdi Rúzsa"
},
{
"operation": "Now on",
"time": "24.292",
"value": "Magdi Rúzsa"
},
{
"operation": "Highlighting",
"time": "48.006",
"value": "Magdi Rúzsa_sentence_3"
},
{
"operation": "hyperlink",
"time": "114.967",
"value": "Fonogram Hungarian Music Awards"
},
{
"operation": "Now on",
"time": "115.572",
"value": "Hungarian Music Awards"
},
{
"operation": "back-button-clicked",
"time": "135.336",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "135.451",
"value": "Magdi Rúzsa"
},
{
"operation": "hyperlink",
"time": "185.422",
"value": "Fonogram Hungarian Music Awards"
},
{
"operation": "Now on",
"time": "185.89",
"value": "Hungarian Music Awards"
},
{
"operation": "search",
"time": "213.411",
"value": "2007 hungarian music awards"
},
{
"operation": "Now on",
"time": "215.075",
"value": "undefined"
},
{
"operation": "finish",
"time": "270.388",
"value": "finish"
}
]
}
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.