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 |
|---|---|---|---|---|---|---|---|---|
60221
|
Sierra de Cazorla (mountain range of the Prebaetic System) is located in Spain with an elevation of 1,847 meters and it's peak is Gilillo.
|
{
"evidence": [
"Sierra de Cazorla_sentence_2",
"Sierra de Cazorla_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:\nSierra de Cazorla (mountain range of the Prebaetic System) is located in Spain with an elevation of 1,847 meters and it's peak is Gilillo.\n",
"role": "user"
}
] |
Entity Disambiguation
|
{
"ground_truth": {
"evidence": [
"Sierra de Cazorla_sentence_2",
"Sierra de Cazorla_sentence_4"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 4684,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.44",
"value": "?search="
},
{
"operation": "search",
"time": "32.571",
"value": "Sierra de Cazoria"
},
{
"operation": "Now on",
"time": "34.058",
"value": "undefined"
},
{
"operation": "hyperlink",
"time": "47.574",
"value": "Sierras de Cazorla, Segura y Las Villas Natural Park"
},
{
"operation": "Now on",
"time": "48.723",
"value": "Sierras de Cazorla, Segura y Las Villas Natural Park"
},
{
"operation": "Highlighting",
"time": "109.381",
"value": "Sierras de Cazorla, Segura y Las Villas Natural Park_sentence_6"
},
{
"operation": "Highlighting",
"time": "109.383",
"value": "Sierras de Cazorla, Segura y Las Villas Natural Park_title"
},
{
"operation": "Highlighting",
"time": "109.385",
"value": "Sierras de Cazorla, Segura y Las Villas Natural Park_section_0"
},
{
"operation": "hyperlink",
"time": "115.122",
"value": "Sierra de Cazorla"
},
{
"operation": "hyperlink",
"time": "118.879",
"value": "Sierra de Cazorla"
},
{
"operation": "Now on",
"time": "132.474",
"value": "Sierra de Cazorla"
},
{
"operation": "Highlighting",
"time": "142.503",
"value": "Sierra de Cazorla_sentence_2"
},
{
"operation": "Highlighting",
"time": "172.347",
"value": "Sierra de Cazorla_sentence_4"
},
{
"operation": "finish",
"time": "247.082",
"value": "finish"
}
]
}
|
exfever_7886
|
The Belko Experiment is a horror-thriller film starring an actor who also starred in a drama series by an Asian playwright and screenwriter.
|
{
"evidence": [
"The Belko Experiment",
"John Gallagher Jr.",
"Aaron Sorkin"
],
"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 Belko Experiment is a horror-thriller film starring an actor who also starred in a drama series by an Asian playwright and screenwriter.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"The Belko Experiment",
"John Gallagher Jr.",
"Aaron Sorkin"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 4685,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_4776
|
Steven Paul Jobs was the wealthiest American inventor in 2011 before dying.
|
{
"evidence": [
"Steve Jobs",
"Steve Wozniak"
],
"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:\nSteven Paul Jobs was the wealthiest American inventor in 2011 before dying.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Steve Jobs",
"Steve Wozniak"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 4686,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
83497
|
Shahodra is a village in India with a total population made up of 54 male and 59 female, as of 2011.
|
{
"evidence": [
"Ashoknagar district_sentence_9",
"Ashoknagar district_sentence_11"
],
"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:\nShahodra is a village in India with a total population made up of 54 male and 59 female, as of 2011.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Ashoknagar district_sentence_9",
"Ashoknagar district_sentence_11"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 4687,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.839",
"value": "?search="
},
{
"operation": "search",
"time": "5.703",
"value": "Shadora"
},
{
"operation": "Now on",
"time": "6.303",
"value": "undefined"
},
{
"operation": "hyperlink",
"time": "15.036",
"value": "Bamori (Vidhan Sabha constituency)"
},
{
"operation": "Now on",
"time": "15.888",
"value": "Bamori (Vidhan Sabha constituency)"
},
{
"operation": "back-button-clicked",
"time": "44.389",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "44.918",
"value": "undefined"
},
{
"operation": "hyperlink",
"time": "46.56",
"value": "Ashoknagar district"
},
{
"operation": "Now on",
"time": "47.673",
"value": "Ashoknagar district"
},
{
"operation": "Highlighting",
"time": "55.582",
"value": "Ashoknagar district_sentence_9"
},
{
"operation": "hyperlink",
"time": "59.969",
"value": "tehsils"
},
{
"operation": "Now on",
"time": "60.594",
"value": "Tehsil"
},
{
"operation": "Highlighting",
"time": "64.481",
"value": "Tehsil_sentence_2"
},
{
"operation": "Highlighting",
"time": "64.481",
"value": "Tehsil_title"
},
{
"operation": "hyperlink",
"time": "72.873",
"value": "List of subdistricts in India"
},
{
"operation": "Now on",
"time": "73.683",
"value": "List of subdistricts in India"
},
{
"operation": "search",
"time": "109.53",
"value": "list of villages in India"
},
{
"operation": "Now on",
"time": "114.951",
"value": "undefined"
},
{
"operation": "hyperlink",
"time": "121.71",
"value": "List of villages Amer, India_section_2"
},
{
"operation": "Now on",
"time": "122.558",
"value": "Amer, India#List of villages Amer.2C India section 2"
},
{
"operation": "back-button-clicked",
"time": "132.173",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "133.306",
"value": "undefined"
},
{
"operation": "back-button-clicked",
"time": "134.378",
"value": "back-button-clicked"
},
{
"operation": "back-button-clicked",
"time": "134.383",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "134.53",
"value": "List of subdistricts in India"
},
{
"operation": "back-button-clicked",
"time": "147.586",
"value": "back-button-clicked"
},
{
"operation": "back-button-clicked",
"time": "147.589",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "147.778",
"value": "Tehsil"
},
{
"operation": "back-button-clicked",
"time": "152.626",
"value": "back-button-clicked"
},
{
"operation": "back-button-clicked",
"time": "152.63",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "153.047",
"value": "Ashoknagar district"
},
{
"operation": "Highlighting",
"time": "155.831",
"value": "Ashoknagar district_sentence_11"
},
{
"operation": "hyperlink",
"time": "188.367",
"value": "2011 census"
},
{
"operation": "Now on",
"time": "190.4",
"value": "2011 Census of India"
},
{
"operation": "hyperlink",
"time": "300.44",
"value": "Demographics of India"
},
{
"operation": "Now on",
"time": "304.536",
"value": "Demographics of India"
},
{
"operation": "hyperlink",
"time": "317.079",
"value": "India"
},
{
"operation": "Now on",
"time": "319.289",
"value": "India"
},
{
"operation": "hyperlink",
"time": "330.422",
"value": "Outline of India"
},
{
"operation": "Now on",
"time": "332.489",
"value": "Outline of India"
},
{
"operation": "back-button-clicked",
"time": "344.98",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "345.228",
"value": "India"
},
{
"operation": "back-button-clicked",
"time": "345.587",
"value": "back-button-clicked"
},
{
"operation": "back-button-clicked",
"time": "345.592",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "346.098",
"value": "Demographics of India"
},
{
"operation": "back-button-clicked",
"time": "346.442",
"value": "back-button-clicked"
},
{
"operation": "back-button-clicked",
"time": "346.447",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "346.673",
"value": "2011 Census of India"
},
{
"operation": "back-button-clicked",
"time": "347.174",
"value": "back-button-clicked"
},
{
"operation": "back-button-clicked",
"time": "347.178",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "347.325",
"value": "Ashoknagar district"
},
{
"operation": "back-button-clicked",
"time": "353.26",
"value": "back-button-clicked"
},
{
"operation": "back-button-clicked",
"time": "353.265",
"value": "back-button-clicked"
},
{
"operation": "back-button-clicked",
"time": "353.268",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "354.699",
"value": "undefined"
},
{
"operation": "hyperlink",
"time": "359.233",
"value": "Bamori (Vidhan Sabha constituency)"
},
{
"operation": "Now on",
"time": "359.719",
"value": "Bamori (Vidhan Sabha constituency)"
},
{
"operation": "back-button-clicked",
"time": "369.629",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "370.16",
"value": "undefined"
},
{
"operation": "hyperlink",
"time": "382.026",
"value": "Bamori (Vidhan Sabha constituency)"
},
{
"operation": "Now on",
"time": "382.788",
"value": "Bamori (Vidhan Sabha constituency)"
},
{
"operation": "hyperlink",
"time": "395.503",
"value": "delimitation"
},
{
"operation": "Now on",
"time": "396.221",
"value": "Boundary delimitation"
},
{
"operation": "back-button-clicked",
"time": "403.098",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "403.195",
"value": "Bamori (Vidhan Sabha constituency)"
},
{
"operation": "back-button-clicked",
"time": "408.633",
"value": "back-button-clicked"
},
{
"operation": "back-button-clicked",
"time": "408.637",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "409.422",
"value": "undefined"
},
{
"operation": "hyperlink",
"time": "411.175",
"value": "Ashoknagar district"
},
{
"operation": "Now on",
"time": "411.768",
"value": "Ashoknagar district"
},
{
"operation": "finish",
"time": "474.031",
"value": "finish"
}
]
}
|
exfever_6822
|
Daft Punk is the first musical duo that introduce techno-pop in the music industry.
|
{
"evidence": [
"Guy-Manuel de Homem-Christo",
"French house"
],
"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:\nDaft Punk is the first musical duo that introduce techno-pop in the music industry.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Guy-Manuel de Homem-Christo",
"French house"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 4688,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_1849
|
Ann Romney is a philanthropist who was also the First Lady of Massachusetts, while her husband served as the head of the state cabinet and the secretary of the commonwealth's military forces.
|
{
"evidence": [
"Ann Romney",
"Governor of Massachusetts"
],
"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:\nAnn Romney is a philanthropist who was also the First Lady of Massachusetts, while her husband served as the head of the state cabinet and the secretary of the commonwealth's military forces.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Ann Romney",
"Governor of Massachusetts"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 4689,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
85046
|
The thin solar panels placed on golf carts convert the sun's energy into charge for the cart, which takes it off the electric grid, and increases drivign distance and battery life; they are widely used by owners who drive long distances, or have fleets of carts.
|
{
"evidence": [
"Solar golf cart_sentence_0",
"Solar golf cart_sentence_1",
"Solar golf cart_sentence_2",
"Solar golf cart_sentence_7"
],
"label": "SUPPORT"
}
|
feverous
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nThe thin solar panels placed on golf carts convert the sun's energy into charge for the cart, which takes it off the electric grid, and increases drivign distance and battery life; they are widely used by owners who drive long distances, or have fleets of carts.\n",
"role": "user"
}
] |
Entity Disambiguation
|
{
"ground_truth": {
"evidence": [
"Solar golf cart_sentence_0",
"Solar golf cart_sentence_1",
"Solar golf cart_sentence_2",
"Solar golf cart_sentence_7"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 4690,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.854",
"value": "?search="
},
{
"operation": "search",
"time": "24.021",
"value": "Electric golf cart"
},
{
"operation": "Now on",
"time": "25.061",
"value": "Golf cart"
},
{
"operation": "Page search",
"time": "42.514",
"value": "solar panels"
},
{
"operation": "search",
"time": "129.04",
"value": "Contains...thin solar panels placed on golf carts"
},
{
"operation": "Now on",
"time": "130.561",
"value": "undefined"
},
{
"operation": "hyperlink",
"time": "144.437",
"value": "Solar cells Honda_section_12"
},
{
"operation": "Now on",
"time": "146.19",
"value": "Honda#Solar cells Honda section 12"
},
{
"operation": "search",
"time": "169.618",
"value": "Contains...golf carts"
},
{
"operation": "Now on",
"time": "171.179",
"value": "undefined"
},
{
"operation": "hyperlink",
"time": "200.105",
"value": "Solar golf cart"
},
{
"operation": "Now on",
"time": "200.855",
"value": "Solar golf cart"
},
{
"operation": "Highlighting",
"time": "214.732",
"value": "Solar golf cart_sentence_0"
},
{
"operation": "Highlighting",
"time": "216.643",
"value": "Solar golf cart_sentence_1"
},
{
"operation": "Highlighting",
"time": "222.361",
"value": "Solar golf cart_sentence_2"
},
{
"operation": "Highlighting",
"time": "248.527",
"value": "Solar golf cart_sentence_7"
},
{
"operation": "finish",
"time": "298.487",
"value": "finish"
}
]
}
|
exfever_2924
|
Warren Casey was an American dancer who writes lyrics for songs.
|
{
"evidence": [
"Warren Casey",
"Lyricist"
],
"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:\nWarren Casey was an American dancer who writes lyrics for songs.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Warren Casey",
"Lyricist"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 4691,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_3493
|
Karl Marx's notable ideas include Marxism.
|
{
"evidence": [
"Law of value",
"The Poverty of Philosophy",
"Anarchism"
],
"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:\nKarl Marx's notable ideas include Marxism.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Law of value",
"The Poverty of Philosophy",
"Anarchism"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 4692,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_6564
|
Escape from Planet Earth is a 2013 3D computer-animated comedy-adventure film genre that uses speculative depictions of phenomena that are not fully accepted by mainstream science.
|
{
"evidence": [
"Escape from Planet Earth",
"Science fiction 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:\nEscape from Planet Earth is a 2013 3D computer-animated comedy-adventure film genre that uses speculative depictions of phenomena that are not fully accepted by mainstream science.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Escape from Planet Earth",
"Science fiction film"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 4693,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_1764
|
Whitney Elizabeth Houston becomes popular with her loud singing voice that reaches every high note of the music she is singing.
|
{
"evidence": [
"Whitney Houston",
"Popular 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:\nWhitney Elizabeth Houston becomes popular with her loud singing voice that reaches every high note of the music she is singing.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Whitney Houston",
"Popular music"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 4694,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_2368
|
Kyra Sedgwick is the youngest American actress who won a Golden Globe Award in 2007.
|
{
"evidence": [
"Kyra Sedgwick",
"Brenda Leigh Johnson"
],
"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:\nKyra Sedgwick is the youngest American actress who won a Golden Globe Award in 2007.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Kyra Sedgwick",
"Brenda Leigh Johnson"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 4695,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
28729
|
In the 1990s, 16 top players left the BDO and formed the Professional Darts Corporation after popularity for the sport hit a low.
|
{
"evidence": [
"Professional Darts Corporation_sentence_7",
"Professional Darts Corporation_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:\nIn the 1990s, 16 top players left the BDO and formed the Professional Darts Corporation after popularity for the sport hit a low.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Professional Darts Corporation_sentence_7",
"Professional Darts Corporation_sentence_8"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 4696,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.039",
"value": "?search="
},
{
"operation": "search",
"time": "16.997",
"value": "Contains...Split in darts"
},
{
"operation": "Now on",
"time": "18.497",
"value": "undefined"
},
{
"operation": "search",
"time": "32.81",
"value": "Professional Darts Corporation"
},
{
"operation": "Now on",
"time": "35.347",
"value": "Professional Darts Corporation"
},
{
"operation": "Page search",
"time": "46.619",
"value": "bdo"
},
{
"operation": "Highlighting",
"time": "49.735",
"value": "Professional Darts Corporation_sentence_8"
},
{
"operation": "Highlighting deleted",
"time": "54.386",
"value": "Professional Darts Corporation_sentence_8"
},
{
"operation": "Highlighting",
"time": "59.6",
"value": "Professional Darts Corporation_sentence_7"
},
{
"operation": "Highlighting",
"time": "67.322",
"value": "Professional Darts Corporation_sentence_8"
},
{
"operation": "finish",
"time": "132.195",
"value": "finish"
}
]
}
|
exfever_1056
|
The law of value is a central concept in Karl Marx's critique of political economy, first expounded in his polemic book that was published in Paris and Brussels in 1837.
|
{
"evidence": [
"Law of value",
"The Poverty of Philosophy"
],
"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 law of value is a central concept in Karl Marx's critique of political economy, first expounded in his polemic book that was published in Paris and Brussels in 1837.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Law of value",
"The Poverty of Philosophy"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 4697,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_4497
|
Amy Adams' first major role came in Catch Me If You Can, by an American film director, producer, and screenwriter who is the recipient of two Academy Awards and has exhibited outstanding directing.
|
{
"evidence": [
"Amy Adams",
"Steven Spielberg",
"Academy Award for Best Director"
],
"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:\nAmy Adams' first major role came in Catch Me If You Can, by an American film director, producer, and screenwriter who is the recipient of two Academy Awards and has exhibited outstanding directing.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Amy Adams",
"Steven Spielberg",
"Academy Award for Best Director"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 4698,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
10253
|
The title Baron de Longueuil was granted to Charles Le Moyne in 1657, and when he died in Montreal in 1729, the barony was passed to his son.
|
{
"evidence": [
"Charles le Moyne de Longueuil, Baron de Longueuil_sentence_0"
],
"label": "NOT ENOUGH INFO"
}
|
feverous
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nThe title Baron de Longueuil was granted to Charles Le Moyne in 1657, and when he died in Montreal in 1729, the barony was passed to his son.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Charles le Moyne de Longueuil, Baron de Longueuil_sentence_0"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 4699,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.079",
"value": "?search="
},
{
"operation": "search",
"time": "9.644",
"value": "charles le moyne"
},
{
"operation": "Now on",
"time": "10.405",
"value": "Charles Le Moyne"
},
{
"operation": "hyperlink",
"time": "19.7",
"value": "Charles le Moyne de Longueuil, Baron de Longueuil"
},
{
"operation": "Now on",
"time": "20.282",
"value": "Charles le Moyne de Longueuil, Baron de Longueuil"
},
{
"operation": "Highlighting",
"time": "25.908",
"value": "Charles le Moyne de Longueuil, Baron de Longueuil_sentence_0"
},
{
"operation": "search",
"time": "78.084",
"value": "baron du longueuil"
},
{
"operation": "Now on",
"time": "80.399",
"value": "undefined"
},
{
"operation": "hyperlink",
"time": "98.139",
"value": "Beloeil, Quebec"
},
{
"operation": "Now on",
"time": "99.06",
"value": "Beloeil, Quebec"
},
{
"operation": "back-button-clicked",
"time": "127.748",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "128.507",
"value": "undefined"
},
{
"operation": "back-button-clicked",
"time": "136.676",
"value": "back-button-clicked"
},
{
"operation": "back-button-clicked",
"time": "136.681",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "136.892",
"value": "Charles le Moyne de Longueuil, Baron de Longueuil"
},
{
"operation": "finish",
"time": "185.804",
"value": "finish"
}
]
}
|
exfever_4521
|
Alfred Thomas Highmore rose to prominence as a youngster in the films Finding Neverland and a musical fantasy film directed by Tim Burton and written by John August, based on Roald Dahl's 1964 British novel of the same name.
|
{
"evidence": [
"Freddie Highmore",
"Charlie and the Chocolate Factory (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:\nAlfred Thomas Highmore rose to prominence as a youngster in the films Finding Neverland and a musical fantasy film directed by Tim Burton and written by John August, based on Roald Dahl's 1964 British novel of the same name.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Freddie Highmore",
"Charlie and the Chocolate Factory (film)"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 4700,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
47219
|
The length of the Emmarentia dam is about 400 meters (1,310 ft.) the whole area surrounding it allows for extensive walking, and an extensive variety of trees, birdlife, and small mammals can be seen.
|
{
"evidence": [
"Emmarentia Dam_sentence_12",
"Emmarentia Dam_sentence_15",
"Emmarentia Dam_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:\nThe length of the Emmarentia dam is about 400 meters (1,310 ft.) the whole area surrounding it allows for extensive walking, and an extensive variety of trees, birdlife, and small mammals can be seen.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Emmarentia Dam_sentence_12",
"Emmarentia Dam_sentence_15",
"Emmarentia Dam_sentence_17"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 4701,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.899",
"value": "?search="
},
{
"operation": "search",
"time": "14.211",
"value": "Emmarentia Dam"
},
{
"operation": "Now on",
"time": "15.509",
"value": "Emmarentia Dam"
},
{
"operation": "Highlighting",
"time": "33.616",
"value": "Emmarentia Dam_sentence_12"
},
{
"operation": "Highlighting",
"time": "40.197",
"value": "Emmarentia Dam_sentence_15"
},
{
"operation": "Highlighting",
"time": "45.713",
"value": "Emmarentia Dam_sentence_17"
},
{
"operation": "finish",
"time": "73.656",
"value": "finish"
}
]
}
|
exfever_7891
|
Willie Hugh Nelson is an American musician, actor, and activist who has made several albums that were released in April 1975, including ten songs that consist entirely of pop standards that Nelson picked from among his favorites.
|
{
"evidence": [
"Willie Nelson",
"Stardust (Willie Nelson 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:\nWillie Hugh Nelson is an American musician, actor, and activist who has made several albums that were released in April 1975, including ten songs that consist entirely of pop standards that Nelson picked from among his favorites.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Willie Nelson",
"Stardust (Willie Nelson album)"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 4702,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
77873
|
In the 1997 Conference, an Alabama school club was with the D-II Independent, while in the 1998 Conference, it was changed to MIFC.
|
{
"evidence": [
"NCAA Division II independent schools_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 1997 Conference, an Alabama school club was with the D-II Independent, while in the 1998 Conference, it was changed to MIFC.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"NCAA Division II independent schools_sentence_0"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 4703,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.873",
"value": "?search="
},
{
"operation": "search",
"time": "11.914",
"value": "1997 Conference alabama"
},
{
"operation": "Now on",
"time": "14.113",
"value": "undefined"
},
{
"operation": "hyperlink",
"time": "20.661",
"value": "Southwestern Athletic Conference"
},
{
"operation": "Now on",
"time": "24.63",
"value": "Southwestern Athletic Conference"
},
{
"operation": "search",
"time": "52.011",
"value": "Alabama school club was with the D-II Independent"
},
{
"operation": "Now on",
"time": "55.104",
"value": "undefined"
},
{
"operation": "search",
"time": "64.566",
"value": "1997 conference D-II"
},
{
"operation": "Now on",
"time": "65.716",
"value": "undefined"
},
{
"operation": "hyperlink",
"time": "72.311",
"value": "NCAA Division II College basketball_section_10"
},
{
"operation": "Now on",
"time": "74.052",
"value": "College basketball#NCAA Division II College basketball section 10"
},
{
"operation": "search",
"time": "110.373",
"value": "1998 Conference USA Men's Basketball Tournament"
},
{
"operation": "Now on",
"time": "111.632",
"value": "1998 Conference USA Men's Basketball Tournament"
},
{
"operation": "Page search",
"time": "117.632",
"value": "MIFC"
},
{
"operation": "hyperlink",
"time": "127.792",
"value": "1998 NCAA Tournament"
},
{
"operation": "Now on",
"time": "130.302",
"value": "1998 NCAA Division I Men's Basketball Tournament"
},
{
"operation": "hyperlink",
"time": "141.596",
"value": "1998 NCAA Division II Men's Basketball Tournament"
},
{
"operation": "Now on",
"time": "142.602",
"value": "1998 NCAA Division II Men's Basketball Tournament"
},
{
"operation": "hyperlink",
"time": "153.349",
"value": "Division II"
},
{
"operation": "Now on",
"time": "154.817",
"value": "NCAA Division II"
},
{
"operation": "search",
"time": "161.328",
"value": "D-II Independent"
},
{
"operation": "Now on",
"time": "163.102",
"value": "undefined"
},
{
"operation": "hyperlink",
"time": "166.877",
"value": "Independents NCAA Division III_section_4"
},
{
"operation": "Now on",
"time": "167.858",
"value": "NCAA Division III#Independents NCAA Division III section 4"
},
{
"operation": "hyperlink",
"time": "177.402",
"value": "Division II"
},
{
"operation": "Now on",
"time": "179.08",
"value": "NCAA Division II"
},
{
"operation": "Page search",
"time": "181.206",
"value": "inde"
},
{
"operation": "hyperlink",
"time": "186.486",
"value": "3.2 Independents NCAA Division II_section_9"
},
{
"operation": "hyperlink",
"time": "189.522",
"value": "Division II Independent schools"
},
{
"operation": "Now on",
"time": "192.004",
"value": "NCAA Division II independent schools"
},
{
"operation": "Highlighting",
"time": "255.699",
"value": "NCAA Division II independent schools_sentence_0"
},
{
"operation": "finish",
"time": "259.318",
"value": "finish"
}
]
}
|
91133
|
Rensselaer Polytechnic Institute's most recen tresearch center is the Center for Biotechnology and Interdisciplinary Studies, located on 15th street between RPI’s Playhouse and Academy Hall which targets biologics, and CBIS has produced over 2,000 peer-reviewed publications.
|
{
"evidence": [
"Rensselaer Polytechnic Institute_sentence_0",
"Center for Biotechnology and Interdisciplinary Studies_sentence_0",
"Center for Biotechnology and Interdisciplinary Studies_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:\nRensselaer Polytechnic Institute's most recen tresearch center is the Center for Biotechnology and Interdisciplinary Studies, located on 15th street between RPI’s Playhouse and Academy Hall which targets biologics, and CBIS has produced over 2,000 peer-reviewed publications.\n",
"role": "user"
}
] |
Multi-hop Reasoning
|
{
"ground_truth": {
"evidence": [
"Rensselaer Polytechnic Institute_sentence_0",
"Center for Biotechnology and Interdisciplinary Studies_sentence_0",
"Center for Biotechnology and Interdisciplinary Studies_sentence_1"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 4704,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.392",
"value": "undefined"
},
{
"operation": "search",
"time": "34.77",
"value": "Rensselaer Polytechnic Institute"
},
{
"operation": "Now on",
"time": "37.209",
"value": "Rensselaer Polytechnic Institute"
},
{
"operation": "Highlighting",
"time": "46.646",
"value": "Rensselaer Polytechnic Institute_sentence_0"
},
{
"operation": "Page search",
"time": "86.554",
"value": "Center for Biotechnology and Interdisciplinary Studies"
},
{
"operation": "Page search",
"time": "86.732",
"value": "Center for Biotechnology and Interdisciplinary Studies"
},
{
"operation": "hyperlink",
"time": "154.108",
"value": "Center for Biotechnology and Interdisciplinary Studies"
},
{
"operation": "Now on",
"time": "155.467",
"value": "Center for Biotechnology and Interdisciplinary Studies"
},
{
"operation": "Highlighting",
"time": "160.199",
"value": "Center for Biotechnology and Interdisciplinary Studies_sentence_0"
},
{
"operation": "Highlighting",
"time": "160.752",
"value": "Center for Biotechnology and Interdisciplinary Studies_sentence_1"
},
{
"operation": "Page search",
"time": "220.358",
"value": "biologics"
},
{
"operation": "Page search",
"time": "220.512",
"value": "biologics"
},
{
"operation": "Page search",
"time": "223.624",
"value": "biologics"
},
{
"operation": "Page search",
"time": "223.823",
"value": "biologics"
},
{
"operation": "Page search",
"time": "224.015",
"value": "biologics"
},
{
"operation": "Page search",
"time": "224.191",
"value": "biologics"
},
{
"operation": "Page search",
"time": "224.368",
"value": "biologics"
},
{
"operation": "search",
"time": "306.771",
"value": "biologics"
},
{
"operation": "Now on",
"time": "308.586",
"value": "Biologics"
},
{
"operation": "back-button-clicked",
"time": "325.911",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "326.55",
"value": "Center for Biotechnology and Interdisciplinary Studies"
},
{
"operation": "Page search",
"time": "333.029",
"value": "biolog"
},
{
"operation": "Page search",
"time": "333.379",
"value": "biolog"
},
{
"operation": "Page search",
"time": "333.579",
"value": "biolog"
},
{
"operation": "Page search",
"time": "333.78",
"value": "biolog"
},
{
"operation": "Page search",
"time": "333.972",
"value": "biolog"
},
{
"operation": "Page search",
"time": "334.163",
"value": "biolog"
},
{
"operation": "Page search",
"time": "334.324",
"value": "biolog"
},
{
"operation": "Page search",
"time": "445.919",
"value": "cbis"
},
{
"operation": "Page search",
"time": "446.384",
"value": "cbis"
},
{
"operation": "Page search",
"time": "446.567",
"value": "cbis"
},
{
"operation": "Page search",
"time": "446.742",
"value": "cbis"
},
{
"operation": "Page search",
"time": "446.918",
"value": "cbis"
},
{
"operation": "Page search",
"time": "447.07",
"value": "cbis"
},
{
"operation": "Page search",
"time": "447.214",
"value": "cbis"
},
{
"operation": "Page search",
"time": "475.76",
"value": "publi"
},
{
"operation": "Page search",
"time": "476.11",
"value": "publi"
},
{
"operation": "Page search",
"time": "476.322",
"value": "publi"
},
{
"operation": "Page search",
"time": "476.535",
"value": "publi"
},
{
"operation": "Page search",
"time": "543.852",
"value": "revi"
},
{
"operation": "Page search",
"time": "544.106",
"value": "revi"
},
{
"operation": "Page search",
"time": "544.323",
"value": "revi"
},
{
"operation": "Page search",
"time": "544.498",
"value": "revi"
},
{
"operation": "Page search",
"time": "544.708",
"value": "revi"
},
{
"operation": "Page search",
"time": "544.898",
"value": "revi"
},
{
"operation": "Page search",
"time": "545.09",
"value": "revi"
},
{
"operation": "Page search",
"time": "545.281",
"value": "revi"
},
{
"operation": "Page search",
"time": "545.49",
"value": "revi"
},
{
"operation": "Page search",
"time": "545.658",
"value": "revi"
},
{
"operation": "finish",
"time": "679.53",
"value": "finish"
}
]
}
|
22254
|
After leaving the CSU, Gabriele Pauli joined the Freie Wähler Bayern which was founded in 2008.
|
{
"evidence": [
"Gabriele Pauli_sentence_6",
"Gabriele Pauli_sentence_7",
"Free Voters_sentence_1"
],
"label": "SUPPORT"
}
|
feverous
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nAfter leaving the CSU, Gabriele Pauli joined the Freie Wähler Bayern which was founded in 2008.\n",
"role": "user"
}
] |
Multi-hop Reasoning
|
{
"ground_truth": {
"evidence": [
"Gabriele Pauli_sentence_6",
"Gabriele Pauli_sentence_7",
"Free Voters_sentence_1"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 4705,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.612",
"value": "?search="
},
{
"operation": "search",
"time": "6.771",
"value": "Gabriele Pauli"
},
{
"operation": "Now on",
"time": "7.637",
"value": "Gabriele Pauli"
},
{
"operation": "Highlighting",
"time": "18.489",
"value": "Gabriele Pauli_sentence_6"
},
{
"operation": "Highlighting",
"time": "20.381",
"value": "Gabriele Pauli_sentence_7"
},
{
"operation": "hyperlink",
"time": "23.095",
"value": "Freie Wähler"
},
{
"operation": "Now on",
"time": "23.951",
"value": "Freie Wähler"
},
{
"operation": "Highlighting",
"time": "29.035",
"value": "Free Voters_header_cell_0_3_0"
},
{
"operation": "Highlighting",
"time": "29.716",
"value": "Free Voters_cell_0_3_1"
},
{
"operation": "Highlighting",
"time": "29.725",
"value": "Free Voters_header_cell_0_3_0"
},
{
"operation": "Highlighting",
"time": "29.725",
"value": "Free Voters_header_cell_0_0_0"
},
{
"operation": "Highlighting deleted",
"time": "36.091",
"value": "Free Voters_cell_0_3_1"
},
{
"operation": "Highlighting deleted",
"time": "36.099",
"value": "Free Voters_header_cell_0_3_0"
},
{
"operation": "Highlighting deleted",
"time": "36.1",
"value": "Free Voters_header_cell_0_0_0"
},
{
"operation": "Highlighting",
"time": "37.583",
"value": "Free Voters_sentence_1"
},
{
"operation": "finish",
"time": "51.34",
"value": "finish"
}
]
}
|
49837
|
Qari Ahmadullah was an educated man according to Abdul Haq and was supposedly killed in the opening days of 2003 in the American airstrike against Mullah Taha's house in Zadran.
|
{
"evidence": [
"Qari Ahmadullah_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:\nQari Ahmadullah was an educated man according to Abdul Haq and was supposedly killed in the opening days of 2003 in the American airstrike against Mullah Taha's house in Zadran.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Qari Ahmadullah_sentence_4"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 4706,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.351",
"value": "?search="
},
{
"operation": "search",
"time": "16.077",
"value": "Qari Ahmadullah"
},
{
"operation": "Now on",
"time": "17.792",
"value": "Qari Ahmadullah"
},
{
"operation": "Highlighting",
"time": "42.931",
"value": "Qari Ahmadullah_sentence_4"
},
{
"operation": "finish",
"time": "61.222",
"value": "finish"
}
]
}
|
exfever_4510
|
The Moscovium was the last synthetic element in the table of elements.
|
{
"evidence": [
"Moscovium",
"Symbol (chemistry)",
"Lead"
],
"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 Moscovium was the last synthetic element in the table of elements.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Moscovium",
"Symbol (chemistry)",
"Lead"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 4707,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_8596
|
The Los Angeles Lakers play their home games at Crypto.com Arena, an arena shared with a professional ice hockey team based in Los Angeles that became one of the seven teams that began to play as part of the 1967 National Hockey League expansion, the largest expansion that was undertaken at one time by an established major sports league.
|
{
"evidence": [
"Los Angeles Lakers",
"Los Angeles Kings",
"1967 NHL expansion"
],
"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 Los Angeles Lakers play their home games at Crypto.com Arena, an arena shared with a professional ice hockey team based in Los Angeles that became one of the seven teams that began to play as part of the 1967 National Hockey League expansion, the largest expansion that was undertaken at one time by an established major sports league.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Los Angeles Lakers",
"Los Angeles Kings",
"1967 NHL expansion"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 4708,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_3017
|
How I Met Your Mother was the first sitcom created by Craig Thomas and Carter Bays for CBS.
|
{
"evidence": [
"How I Met Your Mother",
"20th Television"
],
"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:\nHow I Met Your Mother was the first sitcom created by Craig Thomas and Carter Bays for CBS.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"How I Met Your Mother",
"20th Television"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 4709,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_2079
|
Seven Days in May is a 1964 American political thriller film directed by an American film and television director known for social dramas and action/suspense films.?
|
{
"evidence": [
"Seven Days in May",
"John Frankenheimer"
],
"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:\nSeven Days in May is a 1964 American political thriller film directed by an American film and television director known for social dramas and action/suspense films.?\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Seven Days in May",
"John Frankenheimer"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 4710,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_270
|
Precious is a 2009 American drama film co-produced by an American film director who received two Academy Award nominations, including an award that honors a film director who has exhibited outstanding directing.
|
{
"evidence": [
"Precious (film)",
"Lee Daniels",
"Academy Award for Best Director"
],
"label": "SUPPORT"
}
|
exfever
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nPrecious is a 2009 American drama film co-produced by an American film director who received two Academy Award nominations, including an award that honors a film director who has exhibited outstanding directing.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Precious (film)",
"Lee Daniels",
"Academy Award for Best Director"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 4711,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
74323
|
2010 Fujitsu V8 Supercar Series was planned to have seven rounds between March and December, one of the predicted winners was David Russell (born 4 January 1982).
|
{
"evidence": [
"2010 Fujitsu V8 Supercar Series_sentence_2",
"2010 Fujitsu V8 Supercar Series_sentence_9",
"David Russell (racing driver)_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:\n2010 Fujitsu V8 Supercar Series was planned to have seven rounds between March and December, one of the predicted winners was David Russell (born 4 January 1982).\n",
"role": "user"
}
] |
Multi-hop Reasoning
|
{
"ground_truth": {
"evidence": [
"2010 Fujitsu V8 Supercar Series_sentence_2",
"2010 Fujitsu V8 Supercar Series_sentence_9",
"David Russell (racing driver)_sentence_0"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 4712,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.423",
"value": "?search="
},
{
"operation": "search",
"time": "16.319",
"value": "2010 Fujitsu V8 Supercar Series"
},
{
"operation": "Now on",
"time": "28.661",
"value": "2010 Fujitsu V8 Supercar Series"
},
{
"operation": "Highlighting",
"time": "46.228",
"value": "2010 Fujitsu V8 Supercar Series_sentence_2"
},
{
"operation": "Highlighting",
"time": "105.606",
"value": "2010 Fujitsu V8 Supercar Series_sentence_9"
},
{
"operation": "Page search",
"time": "124.63",
"value": "russell"
},
{
"operation": "hyperlink",
"time": "163.161",
"value": "David Russell"
},
{
"operation": "Now on",
"time": "169.646",
"value": "David Russell (racing driver)"
},
{
"operation": "Highlighting",
"time": "171.898",
"value": "David Russell (racing driver)_sentence_0"
},
{
"operation": "hyperlink",
"time": "214.972",
"value": "Fujitsu V8 Supercar Series"
},
{
"operation": "Now on",
"time": "216.032",
"value": "2010 Fujitsu V8 Supercar Series"
},
{
"operation": "Page search",
"time": "220.142",
"value": "predict"
},
{
"operation": "finish",
"time": "241.924",
"value": "finish"
}
]
}
|
exfever_250
|
The Northeast megalopolis also known Bos-Wash corridor is the capital and most populous city of the Commonwealth of Massachusetts in the United States and 20th-most populous city in the country.
|
{
"evidence": [
"Northeast megalopolis",
"Boston"
],
"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 Northeast megalopolis also known Bos-Wash corridor is the capital and most populous city of the Commonwealth of Massachusetts in the United States and 20th-most populous city in the country.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Northeast megalopolis",
"Boston"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 4713,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_2201
|
The Lockheed Martin F-22 Raptor is the first American single-seat developed for USAF.
|
{
"evidence": [
"Lockheed Martin F-22 Raptor",
"United States Air Force"
],
"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 Lockheed Martin F-22 Raptor is the first American single-seat developed for USAF.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Lockheed Martin F-22 Raptor",
"United States Air Force"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 4714,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
13076
|
Among the A roads in Zone 5 of the Great Britain, Road A55 is the least used because it leads to a dead-end.
|
{
"evidence": [
"A55 road_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:\nAmong the A roads in Zone 5 of the Great Britain, Road A55 is the least used because it leads to a dead-end.\n",
"role": "user"
}
] |
Multi-hop Reasoning
|
{
"ground_truth": {
"evidence": [
"A55 road_sentence_0"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 4715,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.633",
"value": "?search="
},
{
"operation": "search",
"time": "18.957",
"value": "Contains...Road A55"
},
{
"operation": "Now on",
"time": "20.23",
"value": "undefined"
},
{
"operation": "search",
"time": "40.884",
"value": "A roads in Zone 5 of the Great Britain numbering scheme"
},
{
"operation": "Now on",
"time": "42.328",
"value": "A roads in Zone 5 of the Great Britain numbering scheme"
},
{
"operation": "Page search",
"time": "52.937",
"value": "a55"
},
{
"operation": "hyperlink",
"time": "135.194",
"value": "A55"
},
{
"operation": "Now on",
"time": "135.986",
"value": "A55 road (Great Britain)"
},
{
"operation": "search",
"time": "211.663",
"value": "Contains...is the least used because it leads to a dead-end"
},
{
"operation": "Now on",
"time": "216.066",
"value": "undefined"
},
{
"operation": "search",
"time": "230.133",
"value": "A55 road"
},
{
"operation": "Now on",
"time": "230.954",
"value": "A55 road"
},
{
"operation": "Highlighting",
"time": "294.905",
"value": "A55 road_sentence_0"
},
{
"operation": "finish",
"time": "311.264",
"value": "finish"
}
]
}
|
47737
|
Apurinic/apyrimidinic (AP) sites (pre-mutagenic lesions that can prevent normal DNA replication) occur frequently in DNA molecules by spontaneous hydrolysis, by DNA damaging agents or by DNA glycosylases that remove specific abnormal bases.
|
{
"evidence": [
"APEX1_sentence_1",
"APEX1_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:\nApurinic/apyrimidinic (AP) sites (pre-mutagenic lesions that can prevent normal DNA replication) occur frequently in DNA molecules by spontaneous hydrolysis, by DNA damaging agents or by DNA glycosylases that remove specific abnormal bases.\n",
"role": "user"
}
] |
Entity Disambiguation
|
{
"ground_truth": {
"evidence": [
"APEX1_sentence_1",
"APEX1_sentence_2"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 4716,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.662",
"value": "undefined"
},
{
"operation": "search",
"time": "17.243",
"value": "Apurinic/apyrimidinic (AP) sites (pre-mutagenic lesions that can prevent normal DNA replication) occur frequently in DNA molecules by spontaneous hydrolysis, by DNA damaging agents or by DNA glycosylases that remove specific abnormal bases."
},
{
"operation": "Now on",
"time": "18.866",
"value": "undefined"
},
{
"operation": "hyperlink",
"time": "23.948",
"value": "APEX1"
},
{
"operation": "Now on",
"time": "24.695",
"value": "APEX1"
},
{
"operation": "Highlighting",
"time": "28.85",
"value": "APEX1_sentence_1"
},
{
"operation": "Highlighting",
"time": "36.297",
"value": "APEX1_sentence_2"
},
{
"operation": "finish",
"time": "50.173",
"value": "finish"
}
]
}
|
exfever_4873
|
Christian Bale is an English actor who gained wider recognition in American Psycho for his portrayal of a wealthy and materialistic young urban professional, a term coined in the early 1980s for a young professional working in a city.
|
{
"evidence": [
"Christian Bale",
"Patrick Bateman",
"Yuppie"
],
"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:\nChristian Bale is an English actor who gained wider recognition in American Psycho for his portrayal of a wealthy and materialistic young urban professional, a term coined in the early 1980s for a young professional working in a city.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Christian Bale",
"Patrick Bateman",
"Yuppie"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 4717,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
73297
|
Richard Flecknoe attended Jesuit School, founded in 1593 by Father Robert Parsons SJ, and continued his studies at Watten.
|
{
"evidence": [
"Richard Flecknoe_sentence_7",
"Colleges of St Omer, Bruges and Liège_sentence_1",
"Richard Flecknoe_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:\nRichard Flecknoe attended Jesuit School, founded in 1593 by Father Robert Parsons SJ, and continued his studies at Watten.\n",
"role": "user"
}
] |
Multi-hop Reasoning
|
{
"ground_truth": {
"evidence": [
"Richard Flecknoe_sentence_7",
"Colleges of St Omer, Bruges and Liège_sentence_1",
"Richard Flecknoe_sentence_8"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 4718,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.266",
"value": "?search="
},
{
"operation": "search",
"time": "16.854",
"value": "Richard Flecknoe"
},
{
"operation": "Now on",
"time": "18.315",
"value": "Richard Flecknoe"
},
{
"operation": "Highlighting",
"time": "83.033",
"value": "Richard Flecknoe_sentence_7"
},
{
"operation": "hyperlink",
"time": "87.935",
"value": "St Omer English Jesuit School"
},
{
"operation": "Now on",
"time": "89.534",
"value": "Colleges of St Omer, Bruges and Liège"
},
{
"operation": "Highlighting",
"time": "92.488",
"value": "Colleges of St Omer, Bruges and Liège_sentence_1"
},
{
"operation": "back-button-clicked",
"time": "97.708",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "97.749",
"value": "Richard Flecknoe"
},
{
"operation": "Highlighting",
"time": "100.503",
"value": "Richard Flecknoe_sentence_8"
},
{
"operation": "finish",
"time": "110.973",
"value": "finish"
}
]
}
|
exfever_4192
|
Tom Welling is an American actor best known for his role as Clark Kent in an American superhero television series developed by writer-producers Alfred Gough and Miles Millar, based on the Marvel Comics character Ironman.
|
{
"evidence": [
"Tom Welling",
"Smallville"
],
"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:\nTom Welling is an American actor best known for his role as Clark Kent in an American superhero television series developed by writer-producers Alfred Gough and Miles Millar, based on the Marvel Comics character Ironman.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Tom Welling",
"Smallville"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 4719,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_790
|
Moana is a 2016 American 3D computer-animated action film with a?story?interwoven with singing and dancing.
|
{
"evidence": [
"Moana (2016 film)",
"Musical 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:\nMoana is a 2016 American 3D computer-animated action film with a?story?interwoven with singing and dancing.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Moana (2016 film)",
"Musical film"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 4720,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_4764
|
Henry V is a 2019 British television film directed by Thea Sharrock and stars Tom Hiddleston as King of England from 1413 until his death in 1422 and is known and celebrated as one of the greatest warrior-kings of medieval England.
|
{
"evidence": [
"Henry V (2012 film)",
"Henry V of England"
],
"label": "REFUTE"
}
|
exfever
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nHenry V is a 2019 British television film directed by Thea Sharrock and stars Tom Hiddleston as King of England from 1413 until his death in 1422 and is known and celebrated as one of the greatest warrior-kings of medieval England.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Henry V (2012 film)",
"Henry V of England"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 4721,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
71810
|
Lines from The Destiny of Nations (a poem by Coleridge, an English poet, literary critic, philosopher and theologian who was born in 1772) was originally published in the December 26, 1797 Morning Post as The Visions of the Maid of Orleans: A Fragment, but he did not finish it fully until mid 1814.
|
{
"evidence": [
"The Destiny of Nations_sentence_11"
],
"label": "REFUTE"
}
|
feverous
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nLines from The Destiny of Nations (a poem by Coleridge, an English poet, literary critic, philosopher and theologian who was born in 1772) was originally published in the December 26, 1797 Morning Post as The Visions of the Maid of Orleans: A Fragment, but he did not finish it fully until mid 1814.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"The Destiny of Nations_sentence_11"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 4722,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.775",
"value": "?search="
},
{
"operation": "search",
"time": "9.918",
"value": "Lines from The Destiny of Nations"
},
{
"operation": "Now on",
"time": "13.548",
"value": "undefined"
},
{
"operation": "search",
"time": "324.197",
"value": "The Destiny of Nations"
},
{
"operation": "Now on",
"time": "325.741",
"value": "The Destiny of Nations"
},
{
"operation": "Highlighting",
"time": "331.871",
"value": "The Destiny of Nations_sentence_0"
},
{
"operation": "hyperlink",
"time": "334.705",
"value": "Samuel Taylor Coleridge"
},
{
"operation": "Now on",
"time": "340.203",
"value": "Samuel Taylor Coleridge"
},
{
"operation": "Highlighting",
"time": "342.636",
"value": "Samuel Taylor Coleridge_sentence_4"
},
{
"operation": "Highlighting",
"time": "342.638",
"value": "Samuel Taylor Coleridge_title"
},
{
"operation": "back-button-clicked",
"time": "343.86",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "343.972",
"value": "The Destiny of Nations"
},
{
"operation": "Highlighting",
"time": "423.831",
"value": "The Destiny of Nations_sentence_9"
},
{
"operation": "Highlighting",
"time": "444.191",
"value": "The Destiny of Nations_sentence_10"
},
{
"operation": "Highlighting deleted",
"time": "452.339",
"value": "The Destiny of Nations_sentence_10"
},
{
"operation": "Highlighting",
"time": "452.707",
"value": "The Destiny of Nations_sentence_11"
},
{
"operation": "Highlighting deleted",
"time": "453.519",
"value": "The Destiny of Nations_sentence_9"
},
{
"operation": "Highlighting deleted",
"time": "455.127",
"value": "The Destiny of Nations_sentence_0"
},
{
"operation": "finish",
"time": "460.696",
"value": "finish"
}
]
}
|
exfever_7532
|
Jason Bateman played the oldest cast member in the Netflix crime drama series Ozark.
|
{
"evidence": [
"Jason Bateman",
"Little House on the Prairie (TV series)",
"Melissa Gilbert"
],
"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:\nJason Bateman played the oldest cast member in the Netflix crime drama series Ozark.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Jason Bateman",
"Little House on the Prairie (TV series)",
"Melissa Gilbert"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 4723,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
50422
|
HNK Rijeka was founded in 1946 and a Yugoslavian defender named Višković played for them.
|
{
"evidence": [
"HNK Rijeka_sentence_0",
"HNK Rijeka_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:\nHNK Rijeka was founded in 1946 and a Yugoslavian defender named Višković played for them.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"HNK Rijeka_sentence_0",
"HNK Rijeka_sentence_4"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 4724,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.107",
"value": "?search="
},
{
"operation": "search",
"time": "6.795",
"value": "HNK Rijeka"
},
{
"operation": "Now on",
"time": "8.909",
"value": "HNK Rijeka"
},
{
"operation": "Highlighting",
"time": "11.38",
"value": "HNK Rijeka_sentence_0"
},
{
"operation": "Highlighting",
"time": "18.014",
"value": "HNK Rijeka_header_cell_0_0_0"
},
{
"operation": "Highlighting",
"time": "19.107",
"value": "HNK Rijeka_cell_0_0_1"
},
{
"operation": "Highlighting",
"time": "19.121",
"value": "HNK Rijeka_header_cell_0_0_0"
},
{
"operation": "Highlighting",
"time": "20.813",
"value": "HNK Rijeka_header_cell_0_3_0"
},
{
"operation": "Highlighting",
"time": "21.276",
"value": "HNK Rijeka_cell_0_3_1"
},
{
"operation": "Highlighting",
"time": "21.289",
"value": "HNK Rijeka_header_cell_0_3_0"
},
{
"operation": "Highlighting deleted",
"time": "39.001",
"value": "HNK Rijeka_cell_0_3_1"
},
{
"operation": "Highlighting deleted",
"time": "39.007",
"value": "HNK Rijeka_header_cell_0_3_0"
},
{
"operation": "Highlighting deleted",
"time": "40.115",
"value": "HNK Rijeka_cell_0_0_1"
},
{
"operation": "Highlighting deleted",
"time": "40.121",
"value": "HNK Rijeka_header_cell_0_0_0"
},
{
"operation": "Highlighting",
"time": "43.065",
"value": "HNK Rijeka_sentence_4"
},
{
"operation": "search",
"time": "101.573",
"value": "Višković"
},
{
"operation": "Now on",
"time": "102.546",
"value": "Višković"
},
{
"operation": "back-button-clicked",
"time": "126.063",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "126.917",
"value": "HNK Rijeka"
},
{
"operation": "finish",
"time": "137.022",
"value": "finish"
}
]
}
|
19710
|
Eldridge and Hughes was featured on the cover of The Advocate magazine's "Forty Under 40" issue in March 2011, then they got married on June 30, 2012, and reside in Shokan, New York.
|
{
"evidence": [
"Sean Eldridge_sentence_22"
],
"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:\nEldridge and Hughes was featured on the cover of The Advocate magazine's \"Forty Under 40\" issue in March 2011, then they got married on June 30, 2012, and reside in Shokan, New York.\n",
"role": "user"
}
] |
Search terms not in claim
|
{
"ground_truth": {
"evidence": [
"Sean Eldridge_sentence_22"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 4725,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.634",
"value": "?search="
},
{
"operation": "search",
"time": "7.798",
"value": "Contains...Eldridge and Hughes"
},
{
"operation": "Now on",
"time": "9.009",
"value": "undefined"
},
{
"operation": "back-button-clicked",
"time": "23.82",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "23.88",
"value": "?search="
},
{
"operation": "search",
"time": "55.627",
"value": "Sean Eldridge"
},
{
"operation": "Now on",
"time": "56.578",
"value": "Sean Eldridge"
},
{
"operation": "Page search",
"time": "70.983",
"value": "Advocate"
},
{
"operation": "Highlighting",
"time": "73.612",
"value": "Sean Eldridge_sentence_22"
},
{
"operation": "Highlighting deleted",
"time": "87.769",
"value": "Sean Eldridge_sentence_22"
},
{
"operation": "Highlighting",
"time": "88.543",
"value": "Sean Eldridge_sentence_23"
},
{
"operation": "Highlighting",
"time": "96.843",
"value": "Sean Eldridge_sentence_24"
},
{
"operation": "Highlighting",
"time": "103.27",
"value": "Sean Eldridge_sentence_22"
},
{
"operation": "Highlighting deleted",
"time": "110.915",
"value": "Sean Eldridge_sentence_23"
},
{
"operation": "Highlighting deleted",
"time": "111.316",
"value": "Sean Eldridge_sentence_24"
},
{
"operation": "finish",
"time": "132.205",
"value": "finish"
}
]
}
|
exfever_1680
|
Paul Pierce is an American former professional baseball player who before his professional career became part of the NCAA Men's Basketball All-American team made up of NCAA basketball players voted the best in the country by a variety of organizations.
|
{
"evidence": [
"Paul Pierce",
"NCAA Men's Basketball All-Americans"
],
"label": "REFUTE"
}
|
exfever
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nPaul Pierce is an American former professional baseball player who before his professional career became part of the NCAA Men's Basketball All-American team made up of NCAA basketball players voted the best in the country by a variety of organizations.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Paul Pierce",
"NCAA Men's Basketball All-Americans"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 4726,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_7194
|
Smriti Shriniwas Mandhana is an American cricketer who plays for the Indian women's squad and was nominated for the ICC Women's Player of the Year with an Irish international cricketer who made her senior debut for the Irish national team in July 2014, at the age of 13.
|
{
"evidence": [
"Smriti Mandhana",
"Gaby Lewis"
],
"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:\nSmriti Shriniwas Mandhana is an American cricketer who plays for the Indian women's squad and was nominated for the ICC Women's Player of the Year with an Irish international cricketer who made her senior debut for the Irish national team in July 2014, at the age of 13.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Smriti Mandhana",
"Gaby Lewis"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 4727,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
48048
|
Abaco Air, a company in the consumer services industry and airlines sector, is founded in 1975 with headquarters in Marsh Harbour, Bahamas (a state of the Lucayan Archipelago).
|
{
"evidence": [
"Abaco Air_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:\nAbaco Air, a company in the consumer services industry and airlines sector, is founded in 1975 with headquarters in Marsh Harbour, Bahamas (a state of the Lucayan Archipelago).\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Abaco Air_sentence_0"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 4728,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.994",
"value": "?search="
},
{
"operation": "search",
"time": "5.577",
"value": "Abaco Air"
},
{
"operation": "Now on",
"time": "6.688",
"value": "Abaco Air"
},
{
"operation": "search",
"time": "28.404",
"value": "Abaco Air"
},
{
"operation": "Now on",
"time": "29.098",
"value": "Abaco Air"
},
{
"operation": "Highlighting",
"time": "48.182",
"value": "Abaco Air_sentence_0"
},
{
"operation": "hyperlink",
"time": "61.097",
"value": "Marsh Harbour"
},
{
"operation": "Now on",
"time": "61.653",
"value": "Marsh Harbour"
},
{
"operation": "back-button-clicked",
"time": "82.395",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "82.51",
"value": "Abaco Air"
},
{
"operation": "search",
"time": "109.291",
"value": "Abaco Airlines"
},
{
"operation": "Now on",
"time": "110.941",
"value": "undefined"
},
{
"operation": "hyperlink",
"time": "117.538",
"value": "List of defunct airlines of the Bahamas"
},
{
"operation": "Now on",
"time": "118.072",
"value": "List of defunct airlines of the Bahamas"
},
{
"operation": "finish",
"time": "138.51",
"value": "finish"
}
]
}
|
exfever_2711
|
Good Will Hunting is a 1997 American psychological drama film starring an American actor who has appeared in various Kevin Smith films, including a 1999 American fantasy comedy novel.
|
{
"evidence": [
"Good Will Hunting",
"Ben Affleck",
"Dogma (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:\nGood Will Hunting is a 1997 American psychological drama film starring an American actor who has appeared in various Kevin Smith films, including a 1999 American fantasy comedy novel.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Good Will Hunting",
"Ben Affleck",
"Dogma (film)"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 4729,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_4950
|
Harry Potter is a film series based on the eponymous novels by J. K. Rowling and is distributed by an Canadian film and entertainment company.
|
{
"evidence": [
"Harry Potter (film series)",
"Warner Bros."
],
"label": "REFUTE"
}
|
exfever
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nHarry Potter is a film series based on the eponymous novels by J. K. Rowling and is distributed by an Canadian film and entertainment company.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Harry Potter (film series)",
"Warner Bros."
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 4730,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_7829
|
Titanium is a song by music producer David Guetta that featured the vocals of an American singer and songwriter who refer to as the Queen of Hip-Hop Rock.
|
{
"evidence": [
"Titanium (song)",
"Mary J. Blige"
],
"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:\nTitanium is a song by music producer David Guetta that featured the vocals of an American singer and songwriter who refer to as the Queen of Hip-Hop Rock.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Titanium (song)",
"Mary J. Blige"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 4731,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_4061
|
Slash is one of the greatest guitarists in history and part of Guns N' Roses.
|
{
"evidence": [
"Slash (musician)",
"Slash's Snakepit"
],
"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:\nSlash is one of the greatest guitarists in history and part of Guns N' Roses.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Slash (musician)",
"Slash's Snakepit"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 4732,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_7249
|
Sunrise is a Korean animation studio that is renowned for critically praised and popular original anime series, such as the Japanese military sci-fi mecha anime series directed by Ryosuke Takahashi.
|
{
"evidence": [
"Sunrise (company)",
"Armored Trooper Votoms"
],
"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:\nSunrise is a Korean animation studio that is renowned for critically praised and popular original anime series, such as the Japanese military sci-fi mecha anime series directed by Ryosuke Takahashi.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Sunrise (company)",
"Armored Trooper Votoms"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 4733,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_8893
|
Bob Barker is an American entertainer who has the most successful act during the 1960s in an American record label owned by the Universal Music Group founded by Berry Gordy Jr. as Tamla Records on January 12, 1959.
|
{
"evidence": [
"Diana Ross",
"Motown"
],
"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:\nBob Barker is an American entertainer who has the most successful act during the 1960s in an American record label owned by the Universal Music Group founded by Berry Gordy Jr. as Tamla Records on January 12, 1959.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Diana Ross",
"Motown"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 4734,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
89512
|
Alaska Commercial Company (known as the Northern Commercial Company from 1901 to 1992) acquired the assets of the Russian-American Company in 1867; then in 1887, an ACC subsidiary was formed to breed blue foxes on Chirikof.
|
{
"evidence": [
"Alaska Commercial Company_sentence_1",
"Alaska Commercial Company_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:\nAlaska Commercial Company (known as the Northern Commercial Company from 1901 to 1992) acquired the assets of the Russian-American Company in 1867; then in 1887, an ACC subsidiary was formed to breed blue foxes on Chirikof.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Alaska Commercial Company_sentence_1",
"Alaska Commercial Company_sentence_3"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 4735,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.445",
"value": "?search="
},
{
"operation": "search",
"time": "92.462",
"value": "Alaska Commercial Company"
},
{
"operation": "Now on",
"time": "94.141",
"value": "Alaska Commercial Company"
},
{
"operation": "Highlighting",
"time": "98.472",
"value": "Alaska Commercial Company_sentence_1"
},
{
"operation": "Highlighting",
"time": "110.157",
"value": "Alaska Commercial Company_sentence_3"
},
{
"operation": "hyperlink",
"time": "269.801",
"value": "Maritime Fur Trade"
},
{
"operation": "Now on",
"time": "270.842",
"value": "Maritime fur trade"
},
{
"operation": "finish",
"time": "387.731",
"value": "finish"
}
]
}
|
80681
|
The Order of the Golden Fleece is closely associated with the Holy Roman Empire, but their is a splinter group that adheres to the newer statues of the order.
|
{
"evidence": [
"Order of the Golden Fleece_sentence_43",
"Order of the Golden Fleece_sentence_44",
"Order of the Golden Fleece_sentence_45"
],
"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 Order of the Golden Fleece is closely associated with the Holy Roman Empire, but their is a splinter group that adheres to the newer statues of the order.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Order of the Golden Fleece_sentence_43",
"Order of the Golden Fleece_sentence_44",
"Order of the Golden Fleece_sentence_45"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 4736,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.38",
"value": "?search="
},
{
"operation": "search",
"time": "12.191",
"value": "Order of the Golden Fleece"
},
{
"operation": "Now on",
"time": "13.194",
"value": "Order of the Golden Fleece"
},
{
"operation": "Highlighting",
"time": "61.308",
"value": "Order of the Golden Fleece_sentence_43"
},
{
"operation": "Highlighting",
"time": "118.281",
"value": "Order of the Golden Fleece_sentence_44"
},
{
"operation": "Highlighting",
"time": "120.802",
"value": "Order of the Golden Fleece_sentence_45"
},
{
"operation": "finish",
"time": "152.229",
"value": "finish"
}
]
}
|
69387
|
The 2015 Ironman World Championship was organized by Ironman Copenhagen, a sports event promotion company owned by conglomerate Advance Publications, that produces the Ironman Triathlon, Ironman 70.3, the 5150 series of triathlon races, and other sports events.
|
{
"evidence": [
"2015 Ironman World Championship_sentence_2",
"World Triathlon Corporation_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 2015 Ironman World Championship was organized by Ironman Copenhagen, a sports event promotion company owned by conglomerate Advance Publications, that produces the Ironman Triathlon, Ironman 70.3, the 5150 series of triathlon races, and other sports events.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"2015 Ironman World Championship_sentence_2",
"World Triathlon Corporation_sentence_0"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 4737,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.567",
"value": "?search="
},
{
"operation": "search",
"time": "10.549",
"value": "2015 Ironman World Championship"
},
{
"operation": "Now on",
"time": "11.692",
"value": "2015 Ironman World Championship"
},
{
"operation": "Highlighting",
"time": "15.217",
"value": "2015 Ironman World Championship_sentence_2"
},
{
"operation": "hyperlink",
"time": "37.692",
"value": "World Triathlon Corporation"
},
{
"operation": "Now on",
"time": "38.545",
"value": "World Triathlon Corporation"
},
{
"operation": "Highlighting",
"time": "47.206",
"value": "World Triathlon Corporation_sentence_0"
},
{
"operation": "back-button-clicked",
"time": "55.361",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "55.549",
"value": "2015 Ironman World Championship"
},
{
"operation": "finish",
"time": "63.871",
"value": "finish"
}
]
}
|
66288
|
Koolivayal can be accessed from Mananthavady, which is connected to Periya ghat road, and Kannur, which is also connected to Palchuram mountain road.
|
{
"evidence": [
"Koolivayal_sentence_2",
"Koolivayal_sentence_3",
"Koolivayal_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:\nKoolivayal can be accessed from Mananthavady, which is connected to Periya ghat road, and Kannur, which is also connected to Palchuram mountain road.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Koolivayal_sentence_2",
"Koolivayal_sentence_3",
"Koolivayal_sentence_6"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 4738,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "87.882",
"value": "Keith Wallen#Early years and Copper Keith Wallen section 1"
},
{
"operation": "search",
"time": "91.397",
"value": "Koolivayal can be accessed from Mananthavady"
},
{
"operation": "Now on",
"time": "92.878",
"value": "undefined"
},
{
"operation": "hyperlink",
"time": "96.61",
"value": "Important buildings Koolivayal_section_0"
},
{
"operation": "Now on",
"time": "97.534",
"value": "Koolivayal#Important buildings Koolivayal section 0"
},
{
"operation": "Highlighting",
"time": "100.425",
"value": "Koolivayal_sentence_2"
},
{
"operation": "Highlighting",
"time": "106.168",
"value": "Koolivayal_sentence_3"
},
{
"operation": "Highlighting",
"time": "125.48",
"value": "Koolivayal_sentence_6"
},
{
"operation": "finish",
"time": "129.698",
"value": "finish"
}
]
}
|
exfever_5968
|
SpaceX ended the United Launch Alliance monopoly of U.S. military payloads.
|
{
"evidence": [
"SpaceX",
"Falcon Heavy",
"Launch vehicle"
],
"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:\nSpaceX ended the United Launch Alliance monopoly of U.S. military payloads.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"SpaceX",
"Falcon Heavy",
"Launch vehicle"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 4739,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
23267
|
Tommy Molloy turned professional in February 1955, he was beaten but beat Newman on points to become British champion.
|
{
"evidence": [
"Tommy Molloy_sentence_3",
"Tommy Molloy_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:\nTommy Molloy turned professional in February 1955, he was beaten but beat Newman on points to become British champion.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Tommy Molloy_sentence_3",
"Tommy Molloy_sentence_4"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 4740,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.486",
"value": "?search="
},
{
"operation": "search",
"time": "7.03",
"value": "Tommy Molloy"
},
{
"operation": "Now on",
"time": "7.554",
"value": "Tommy Molloy"
},
{
"operation": "Page search",
"time": "44.551",
"value": "prof"
},
{
"operation": "Highlighting",
"time": "47.173",
"value": "Tommy Molloy_sentence_2"
},
{
"operation": "Page search",
"time": "71.966",
"value": "Newman"
},
{
"operation": "Highlighting",
"time": "103.752",
"value": "Tommy Molloy_sentence_3"
},
{
"operation": "Highlighting",
"time": "104.417",
"value": "Tommy Molloy_sentence_4"
},
{
"operation": "Highlighting deleted",
"time": "105.078",
"value": "Tommy Molloy_sentence_2"
},
{
"operation": "finish",
"time": "189.295",
"value": "finish"
}
]
}
|
exfever_4997
|
Toronto City is the anchor of a secondary region of Southern Ontario, which is part of the most densely populated and heavily industrialized region of the United States.
|
{
"evidence": [
"Toronto",
"Golden Horseshoe",
"Quebec City–Windsor Corridor"
],
"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:\nToronto City is the anchor of a secondary region of Southern Ontario, which is part of the most densely populated and heavily industrialized region of the United States.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Toronto",
"Golden Horseshoe",
"Quebec City–Windsor Corridor"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 4741,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
1202
|
The 2010–11 Cuban National Series, the 50th edition of the tournament, began on Sunday, November 28 with a game between Villa Clara and Industriales.
|
{
"evidence": [
"2010–11 Cuban National Series_sentence_0",
"2010–11 Cuban National Series_sentence_1"
],
"label": "SUPPORT"
}
|
feverous
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nThe 2010–11 Cuban National Series, the 50th edition of the tournament, began on Sunday, November 28 with a game between Villa Clara and Industriales.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"2010–11 Cuban National Series_sentence_0",
"2010–11 Cuban National Series_sentence_1"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 4742,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "search",
"time": "12.999",
"value": "Cuban National Series"
},
{
"operation": "search",
"time": "54.628",
"value": "Cuban National Series 2010"
},
{
"operation": "search",
"time": "75.077",
"value": "2010-11 Cuban National Series"
},
{
"operation": "Highlighting",
"time": "80.235",
"value": "2010–11 Cuban National Series_sentence_0"
},
{
"operation": "Highlighting",
"time": "82.636",
"value": "2010–11 Cuban National Series_sentence_1"
},
{
"operation": "Highlighting",
"time": "99.453",
"value": "2010–11 Cuban National Series_sentence_0"
},
{
"operation": "Highlighting",
"time": "100.645",
"value": "2010–11 Cuban National Series_sentence_1"
},
{
"operation": "search",
"time": "105.624",
"value": "Cuban National Series"
},
{
"operation": "finish",
"time": "148.038",
"value": "finish"
}
]
}
|
82334
|
The Golden Globe Award for Best Actress in a Television Series Musical or Comedy was first presented on May 5, 1962, by the Hollywood Foreign Press Association.
|
{
"evidence": [
"Golden Globe Award for Best Actress – Television Series Musical or Comedy_sentence_2"
],
"label": "REFUTE"
}
|
feverous
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nThe Golden Globe Award for Best Actress in a Television Series Musical or Comedy was first presented on May 5, 1962, by the Hollywood Foreign Press Association.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Golden Globe Award for Best Actress – Television Series Musical or Comedy_sentence_2"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 4743,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.923",
"value": "?search="
},
{
"operation": "search",
"time": "30.691",
"value": "Contains...Golden Globe Award for Best Actress in a Television Series Musical or Comedy"
},
{
"operation": "Now on",
"time": "36.194",
"value": "undefined"
},
{
"operation": "hyperlink",
"time": "42.513",
"value": "Golden Globe Award for Best Actress – Television Series Musical or Comedy"
},
{
"operation": "Now on",
"time": "43.65",
"value": "Golden Globe Award for Best Actress – Television Series Musical or Comedy"
},
{
"operation": "Highlighting",
"time": "65.425",
"value": "Golden Globe Award for Best Actress – Television Series Musical or Comedy_sentence_2"
},
{
"operation": "finish",
"time": "98.278",
"value": "finish"
}
]
}
|
exfever_1288
|
Christopher Carl Carter gained fame in the 1990s as the creator of the Fox American science fiction drama television series that revolves around Federal Bureau of Investigation special agents Fox Mulder and Dana Scully.
|
{
"evidence": [
"Chris Carter (screenwriter)",
"The X-Files"
],
"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:\nChristopher Carl Carter gained fame in the 1990s as the creator of the Fox American science fiction drama television series that revolves around Federal Bureau of Investigation special agents Fox Mulder and Dana Scully.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Chris Carter (screenwriter)",
"The X-Files"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 4744,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_4452
|
Outkast was an American hip hop duo who achieved both critical acclaim and commercial success from the mid-1990s to the early 2000s, helping to popularize a blanket term for a regional genre of American hip hop music.
|
{
"evidence": [
"Outkast",
"Southern hip hop"
],
"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:\nOutkast was an American hip hop duo who achieved both critical acclaim and commercial success from the mid-1990s to the early 2000s, helping to popularize a blanket term for a regional genre of American hip hop music.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Outkast",
"Southern hip hop"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 4745,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_5056
|
Pirates of the Caribbean is a series of fantasy swashbuckler films produced by Jerry Bruckheimer and based on Walt Disney's dark ride tells the story of a band of pirates at a theme park found at Disneyland Canada in Marne-la-Vallée, France.
|
{
"evidence": [
"Pirates of the Caribbean (film series)",
"Pirates of the Caribbean (attraction)",
"Disneyland Park (Paris)"
],
"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:\nPirates of the Caribbean is a series of fantasy swashbuckler films produced by Jerry Bruckheimer and based on Walt Disney's dark ride tells the story of a band of pirates at a theme park found at Disneyland Canada in Marne-la-Vallée, France.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Pirates of the Caribbean (film series)",
"Pirates of the Caribbean (attraction)",
"Disneyland Park (Paris)"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 4746,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
76416
|
Christian Karsten Hansen held positions as senior scientist, co-founded Profound Pharma in Denmark and was involved in investment.
|
{
"evidence": [
"Christian Karsten Hansen_sentence_5",
"Christian Karsten Hansen_sentence_6",
"Christian Karsten Hansen_sentence_7",
"Christian Karsten Hansen_sentence_8",
"Profound Pharma_sentence_1"
],
"label": "SUPPORT"
}
|
feverous
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nChristian Karsten Hansen held positions as senior scientist, co-founded Profound Pharma in Denmark and was involved in investment.\n",
"role": "user"
}
] |
Multi-hop Reasoning
|
{
"ground_truth": {
"evidence": [
"Christian Karsten Hansen_sentence_5",
"Christian Karsten Hansen_sentence_6",
"Christian Karsten Hansen_sentence_7",
"Christian Karsten Hansen_sentence_8",
"Profound Pharma_sentence_1"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 4747,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.63",
"value": "?search="
},
{
"operation": "search",
"time": "20.601",
"value": "Christian Karsten"
},
{
"operation": "Now on",
"time": "21.672",
"value": "undefined"
},
{
"operation": "search",
"time": "34.729",
"value": "Christian Karsten Hansen"
},
{
"operation": "Now on",
"time": "35.245",
"value": "Christian Karsten Hansen"
},
{
"operation": "Highlighting",
"time": "45.089",
"value": "Christian Karsten Hansen_sentence_5"
},
{
"operation": "Highlighting",
"time": "48.544",
"value": "Christian Karsten Hansen_sentence_6"
},
{
"operation": "Highlighting",
"time": "61.791",
"value": "Christian Karsten Hansen_sentence_7"
},
{
"operation": "Highlighting",
"time": "67.127",
"value": "Christian Karsten Hansen_sentence_8"
},
{
"operation": "hyperlink",
"time": "67.927",
"value": "Profound Pharma"
},
{
"operation": "Now on",
"time": "68.326",
"value": "Profound Pharma"
},
{
"operation": "Highlighting",
"time": "71.263",
"value": "Profound Pharma_sentence_1"
},
{
"operation": "finish",
"time": "79.175",
"value": "finish"
}
]
}
|
18292
|
The Pakistani team players outnumbered the Indian cricket team in the test matches and the One Day Internationals in 2007-08.
|
{
"evidence": [
"India national cricket team_sentence_0",
"India national cricket team_sentence_1",
"India national cricket team_sentence_11"
],
"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 Pakistani team players outnumbered the Indian cricket team in the test matches and the One Day Internationals in 2007-08.\n",
"role": "user"
}
] |
Search terms not in claim
|
{
"ground_truth": {
"evidence": [
"India national cricket team_sentence_0",
"India national cricket team_sentence_1",
"India national cricket team_sentence_11"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 4748,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.812",
"value": "?search="
},
{
"operation": "search",
"time": "19.763",
"value": "One Day Internationals 2007-08"
},
{
"operation": "Now on",
"time": "21.142",
"value": "undefined"
},
{
"operation": "hyperlink",
"time": "36.066",
"value": "One-Day International status (2007–present) Ireland cricket team_section_4"
},
{
"operation": "Now on",
"time": "36.947",
"value": "Ireland cricket team#One-Day International status .282007.E2.80.93present.29 Ireland cricket team section 4"
},
{
"operation": "back-button-clicked",
"time": "56.65",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "57.275",
"value": "undefined"
},
{
"operation": "search",
"time": "107.09",
"value": "Pakistani One Day Internationals 2007-08"
},
{
"operation": "Now on",
"time": "108.051",
"value": "undefined"
},
{
"operation": "hyperlink",
"time": "110.177",
"value": "pakistan one day internationals 2007 08"
},
{
"operation": "Now on",
"time": "110.935",
"value": "undefined"
},
{
"operation": "hyperlink",
"time": "120.929",
"value": "One Day Internationals Ajantha Mendis_section_7"
},
{
"operation": "Now on",
"time": "121.502",
"value": "Ajantha Mendis#One Day Internationals Ajantha Mendis section 7"
},
{
"operation": "search",
"time": "149.274",
"value": "One Day Internationals cricket"
},
{
"operation": "Now on",
"time": "150.07",
"value": "undefined"
},
{
"operation": "hyperlink",
"time": "168.825",
"value": "Maharashtra Cricket Association Stadium"
},
{
"operation": "Now on",
"time": "169.314",
"value": "Maharashtra Cricket Association Stadium"
},
{
"operation": "Page search",
"time": "175.25",
"value": "one day internationas"
},
{
"operation": "Page search",
"time": "178.297",
"value": "one day internationals"
},
{
"operation": "hyperlink",
"time": "182.753",
"value": "4.3 One Day Internationals Maharashtra Cricket Association Stadium_section_8"
},
{
"operation": "search",
"time": "220.665",
"value": "test matches one day internationals 2007-08"
},
{
"operation": "Now on",
"time": "221.806",
"value": "undefined"
},
{
"operation": "hyperlink",
"time": "242.281",
"value": "West Indian cricket team in South Africa in 2007–08"
},
{
"operation": "Now on",
"time": "242.726",
"value": "West Indian cricket team in South Africa in 2007–08"
},
{
"operation": "hyperlink",
"time": "270.001",
"value": "2 Test series West Indian cricket team in South Africa in 2007–08_section_3"
},
{
"operation": "hyperlink",
"time": "273.616",
"value": "West Indies cricket team"
},
{
"operation": "Now on",
"time": "274.468",
"value": "West Indies cricket team"
},
{
"operation": "search",
"time": "296.809",
"value": "Indian cricket 2007-08"
},
{
"operation": "Now on",
"time": "297.708",
"value": "undefined"
},
{
"operation": "search",
"time": "322.016",
"value": "Indian Cricket team test matches"
},
{
"operation": "Now on",
"time": "323.324",
"value": "undefined"
},
{
"operation": "hyperlink",
"time": "328.009",
"value": "India national cricket team"
},
{
"operation": "Now on",
"time": "328.909",
"value": "India national cricket team"
},
{
"operation": "hyperlink",
"time": "337.432",
"value": "2007 ICC World Twenty20"
},
{
"operation": "Now on",
"time": "338.007",
"value": "2007 ICC World Twenty20"
},
{
"operation": "back-button-clicked",
"time": "340.656",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "340.846",
"value": "India national cricket team"
},
{
"operation": "Page search",
"time": "349.787",
"value": "one day internationa;s"
},
{
"operation": "Page search",
"time": "351.489",
"value": "one day internationals"
},
{
"operation": "Highlighting",
"time": "362.129",
"value": "India national cricket team_sentence_0"
},
{
"operation": "Highlighting",
"time": "362.743",
"value": "India national cricket team_sentence_1"
},
{
"operation": "Highlighting",
"time": "370.431",
"value": "India national cricket team_sentence_11"
},
{
"operation": "finish",
"time": "378.159",
"value": "finish"
}
]
}
|
47839
|
Jones and Dixon married in 1988, and Jones' sound encompassed jangle pop, ballads, and southern-style soul.
|
{
"evidence": [
"Marti Jones_sentence_13",
"Marti Jones_sentence_10"
],
"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:\nJones and Dixon married in 1988, and Jones' sound encompassed jangle pop, ballads, and southern-style soul.\n",
"role": "user"
}
] |
Entity Disambiguation
|
{
"ground_truth": {
"evidence": [
"Marti Jones_sentence_13",
"Marti Jones_sentence_10"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 4749,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.75",
"value": "?search="
},
{
"operation": "search",
"time": "16.858",
"value": "Jones' sound encompassed jangle pop, ballads, and southern-style soul"
},
{
"operation": "Now on",
"time": "18.036",
"value": "undefined"
},
{
"operation": "hyperlink",
"time": "25.119",
"value": "Marti Jones and Don Dixon Marti Jones_section_9"
},
{
"operation": "Now on",
"time": "25.702",
"value": "Marti Jones#Marti Jones and Don Dixon Marti Jones section 9"
},
{
"operation": "Page search",
"time": "35.651",
"value": "Jones' sound encompassed jangle pop, ballads, and southern-style soul"
},
{
"operation": "Highlighting",
"time": "39.059",
"value": "Marti Jones_sentence_13"
},
{
"operation": "Highlighting",
"time": "43.563",
"value": "Marti Jones_sentence_10"
},
{
"operation": "finish",
"time": "50.828",
"value": "finish"
}
]
}
|
93168
|
Vayalar's secondary religion is Hinduism and Shiva temple, a symbolic house, seat and body of divinity for Hindus, in Nagamkulangara is famous.
|
{
"evidence": [
"Vayalar_sentence_12"
],
"label": "REFUTE"
}
|
feverous
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nVayalar's secondary religion is Hinduism and Shiva temple, a symbolic house, seat and body of divinity for Hindus, in Nagamkulangara is famous.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Vayalar_sentence_12"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 4750,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.85",
"value": "?search="
},
{
"operation": "search",
"time": "10.076",
"value": "Vayalar"
},
{
"operation": "Now on",
"time": "11.116",
"value": "Vayalar"
},
{
"operation": "Page search",
"time": "17.667",
"value": "hindu"
},
{
"operation": "Highlighting",
"time": "21.518",
"value": "Vayalar_sentence_12"
},
{
"operation": "Page search",
"time": "31.386",
"value": "secondary"
},
{
"operation": "Page search",
"time": "35.234",
"value": "second"
},
{
"operation": "finish",
"time": "44.492",
"value": "finish"
}
]
}
|
28066
|
While Piłsudski dodged the bullets, the Lwów Province Governor, Kazimierz Grabowski, was wounded in the opening ceremony of the first Eastern Trade Fair; which by 1928, could boast some 1,600 exhibitors, and about 400 being foreign firms.
|
{
"evidence": [
"Targi Wschodnie_sentence_7",
"Targi Wschodnie_sentence_6",
"Targi Wschodnie_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:\nWhile Piłsudski dodged the bullets, the Lwów Province Governor, Kazimierz Grabowski, was wounded in the opening ceremony of the first Eastern Trade Fair; which by 1928, could boast some 1,600 exhibitors, and about 400 being foreign firms.\n",
"role": "user"
}
] |
Search terms not in claim
|
{
"ground_truth": {
"evidence": [
"Targi Wschodnie_sentence_7",
"Targi Wschodnie_sentence_6",
"Targi Wschodnie_sentence_8"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 4751,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.92",
"value": "?search="
},
{
"operation": "search",
"time": "13.795",
"value": "Targi Wschodnie"
},
{
"operation": "Now on",
"time": "14.712",
"value": "Targi Wschodnie"
},
{
"operation": "Highlighting",
"time": "23.442",
"value": "Targi Wschodnie_sentence_7"
},
{
"operation": "Highlighting",
"time": "30.693",
"value": "Targi Wschodnie_sentence_6"
},
{
"operation": "Highlighting",
"time": "37.051",
"value": "Targi Wschodnie_sentence_8"
},
{
"operation": "finish",
"time": "46.67",
"value": "finish"
}
]
}
|
61511
|
Deep Down Happy is musical studio owned by English alternative rock band Sports Team, "here it comes, again" was the first single released by the band on June 12 2019, and rose to No. 2 on the official charts company's UK Albums Chart. on June 12 2020
|
{
"evidence": [
"Deep Down Happy_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:\nDeep Down Happy is musical studio owned by English alternative rock band Sports Team, \"here it comes, again\" was the first single released by the band on June 12 2019, and rose to No. 2 on the official charts company's UK Albums Chart. on June 12 2020\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Deep Down Happy_sentence_0"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 4752,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.732",
"value": "?search="
},
{
"operation": "search",
"time": "9.097",
"value": "Deep Down Happy"
},
{
"operation": "Now on",
"time": "10.163",
"value": "Deep Down Happy"
},
{
"operation": "Highlighting",
"time": "80.354",
"value": "Deep Down Happy_sentence_0"
},
{
"operation": "finish",
"time": "94.696",
"value": "finish"
}
]
}
|
exfever_1146
|
Ron Dennis is best known as the founder of a British motor racing team.
|
{
"evidence": [
"Ron Dennis",
"McLaren"
],
"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:\nRon Dennis is best known as the founder of a British motor racing team.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Ron Dennis",
"McLaren"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 4753,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
37199
|
The Emergency Banking Act stayed as a draft proposal until today.
|
{
"evidence": [
"Emergency Banking Act_sentence_7"
],
"label": "REFUTE"
}
|
feverous
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nThe Emergency Banking Act stayed as a draft proposal until today.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Emergency Banking Act_sentence_7"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 4754,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.071",
"value": "?search="
},
{
"operation": "search",
"time": "12.397",
"value": "Contains...The Emergency Banking Act"
},
{
"operation": "Now on",
"time": "15.258",
"value": "undefined"
},
{
"operation": "hyperlink",
"time": "20.091",
"value": "Emergency Banking Act"
},
{
"operation": "Now on",
"time": "20.76",
"value": "Emergency Banking Act"
},
{
"operation": "Highlighting",
"time": "34.535",
"value": "Emergency Banking Act_sentence_1"
},
{
"operation": "Page search",
"time": "56.667",
"value": "draft proposal"
},
{
"operation": "Highlighting deleted",
"time": "67.346",
"value": "Emergency Banking Act_sentence_1"
},
{
"operation": "Highlighting",
"time": "70.651",
"value": "Emergency Banking Act_sentence_7"
},
{
"operation": "Highlighting deleted",
"time": "92.403",
"value": "Emergency Banking Act_sentence_7"
},
{
"operation": "Highlighting",
"time": "96.075",
"value": "Emergency Banking Act_sentence_7"
},
{
"operation": "finish",
"time": "111.339",
"value": "finish"
}
]
}
|
exfever_4483
|
Norma Shearer is a dancer.
|
{
"evidence": [
"Joan Crawford",
"Greta Garbo"
],
"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:\nNorma Shearer is a dancer.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Joan Crawford",
"Greta Garbo"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 4755,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
64789
|
Tai Lam Chung Reservoir is the first reservoir with a water storage capacity is about 21 million cubic metres, built in Hong Kong after the Second World War
|
{
"evidence": [
"Tai Lam Chung Reservoir_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:\nTai Lam Chung Reservoir is the first reservoir with a water storage capacity is about 21 million cubic metres, built in Hong Kong after the Second World War\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Tai Lam Chung Reservoir_sentence_4"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 4756,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.841",
"value": "?search="
},
{
"operation": "search",
"time": "26.669",
"value": "Tai Lam Chung Reservoir"
},
{
"operation": "Now on",
"time": "27.91",
"value": "Tai Lam Chung Reservoir"
},
{
"operation": "Highlighting",
"time": "85.547",
"value": "Tai Lam Chung Reservoir_sentence_1"
},
{
"operation": "Highlighting deleted",
"time": "137.267",
"value": "Tai Lam Chung Reservoir_sentence_1"
},
{
"operation": "Highlighting",
"time": "138.371",
"value": "Tai Lam Chung Reservoir_sentence_4"
},
{
"operation": "finish",
"time": "211.931",
"value": "finish"
}
]
}
|
exfever_2044
|
Willow Camille Reign Smith made her acting debut in 2007 in the film I Am Legend along with her father, Will Smith, and later appeared in a 2008 American comedy-drama film directed by Patricia Rozema and written by Ann Peacock.
|
{
"evidence": [
"Willow Smith",
"Kit Kittredge: An American Girl"
],
"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:\nWillow Camille Reign Smith made her acting debut in 2007 in the film I Am Legend along with her father, Will Smith, and later appeared in a 2008 American comedy-drama film directed by Patricia Rozema and written by Ann Peacock.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Willow Smith",
"Kit Kittredge: An American Girl"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 4757,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
37980
|
During the 2009 CD Universidad San Martin season, Luis Álvarez and Pedro Gallese both transferred to Atlético Minero.
|
{
"evidence": [
"Pedro Gallese_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:\nDuring the 2009 CD Universidad San Martin season, Luis Álvarez and Pedro Gallese both transferred to Atlético Minero.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Pedro Gallese_sentence_5"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 4758,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.797",
"value": "?search="
},
{
"operation": "search",
"time": "17.302",
"value": "2009 CD Universidad San Martín season"
},
{
"operation": "Now on",
"time": "19.051",
"value": "2009 CD Universidad San Martín season"
},
{
"operation": "Page search",
"time": "28.978",
"value": "alvarez"
},
{
"operation": "Page search",
"time": "52.07",
"value": "atletico"
},
{
"operation": "Page search",
"time": "67.333",
"value": "atletico"
},
{
"operation": "Page search",
"time": "128.95",
"value": "pedro"
},
{
"operation": "Highlighting",
"time": "148.593",
"value": "2009 CD Universidad San Martín season_cell_1_35_2"
},
{
"operation": "Highlighting",
"time": "148.594",
"value": "2009 CD Universidad San Martín season_title"
},
{
"operation": "Highlighting",
"time": "148.594",
"value": "2009 CD Universidad San Martín season_section_0"
},
{
"operation": "Highlighting",
"time": "148.595",
"value": "2009 CD Universidad San Martín season_section_1"
},
{
"operation": "Highlighting",
"time": "148.608",
"value": "2009 CD Universidad San Martín season_header_cell_1_0_2"
},
{
"operation": "hyperlink",
"time": "150.92",
"value": "Pedro Gallese"
},
{
"operation": "Now on",
"time": "151.855",
"value": "Pedro Gallese"
},
{
"operation": "Page search",
"time": "160.871",
"value": "atletico"
},
{
"operation": "Highlighting",
"time": "190.109",
"value": "Pedro Gallese_sentence_5"
},
{
"operation": "back-button-clicked",
"time": "198.692",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "199.041",
"value": "2009 CD Universidad San Martín season"
},
{
"operation": "Highlighting",
"time": "208.15",
"value": "2009 CD Universidad San Martín season_cell_1_33_2"
},
{
"operation": "Highlighting",
"time": "208.151",
"value": "2009 CD Universidad San Martín season_title"
},
{
"operation": "Highlighting",
"time": "208.151",
"value": "2009 CD Universidad San Martín season_section_0"
},
{
"operation": "Highlighting",
"time": "208.152",
"value": "2009 CD Universidad San Martín season_section_1"
},
{
"operation": "Highlighting",
"time": "208.162",
"value": "2009 CD Universidad San Martín season_header_cell_1_0_2"
},
{
"operation": "hyperlink",
"time": "212.914",
"value": "Peru"
},
{
"operation": "Now on",
"time": "214.591",
"value": "Peru"
},
{
"operation": "back-button-clicked",
"time": "218.297",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "218.666",
"value": "2009 CD Universidad San Martín season"
},
{
"operation": "search",
"time": "230.805",
"value": "luiz alvarez"
},
{
"operation": "Now on",
"time": "232.17",
"value": "undefined"
},
{
"operation": "back-button-clicked",
"time": "258.299",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "258.665",
"value": "2009 CD Universidad San Martín season"
},
{
"operation": "search",
"time": "277.437",
"value": "Atlético Minero"
},
{
"operation": "Now on",
"time": "278.104",
"value": "Atlético Minero"
},
{
"operation": "search",
"time": "297.648",
"value": "2009 atletico minero"
},
{
"operation": "Now on",
"time": "298.607",
"value": "undefined"
},
{
"operation": "back-button-clicked",
"time": "307.267",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "307.441",
"value": "Atlético Minero"
},
{
"operation": "search",
"time": "340.2",
"value": "atletico minero 2009 season"
},
{
"operation": "Now on",
"time": "341.347",
"value": "undefined"
},
{
"operation": "back-button-clicked",
"time": "373.169",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "373.363",
"value": "Atlético Minero"
},
{
"operation": "Highlighting",
"time": "390.943",
"value": "Atlético Minero_sentence_16"
},
{
"operation": "Highlighting",
"time": "390.944",
"value": "Atlético Minero_title"
},
{
"operation": "Highlighting",
"time": "390.945",
"value": "Atlético Minero_section_2"
},
{
"operation": "Highlighting deleted",
"time": "394.317",
"value": "Atlético Minero_sentence_16"
},
{
"operation": "Highlighting deleted",
"time": "394.319",
"value": "Atlético Minero_title"
},
{
"operation": "Highlighting deleted",
"time": "394.32",
"value": "Atlético Minero_section_2"
},
{
"operation": "search",
"time": "400.574",
"value": "Atlético Minero footballer"
},
{
"operation": "Now on",
"time": "401.487",
"value": "undefined"
},
{
"operation": "finish",
"time": "483.608",
"value": "finish"
}
]
}
|
48765
|
The southern terminus of West Virgina Route 214 is southeast of Griffithsville, while the northern terminus is south of South Charleston.
|
{
"evidence": [
"West Virginia Route 214_sentence_3",
"West Virginia Route 214_sentence_4"
],
"label": "SUPPORT"
}
|
feverous
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nThe southern terminus of West Virgina Route 214 is southeast of Griffithsville, while the northern terminus is south of South Charleston.\n",
"role": "user"
}
] |
Entity Disambiguation
|
{
"ground_truth": {
"evidence": [
"West Virginia Route 214_sentence_3",
"West Virginia Route 214_sentence_4"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 4759,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.862",
"value": "?search="
},
{
"operation": "search",
"time": "18.752",
"value": "Contains...West Virgina Route 214"
},
{
"operation": "Now on",
"time": "20.491",
"value": "undefined"
},
{
"operation": "hyperlink",
"time": "35.664",
"value": "west virginia route 214"
},
{
"operation": "Now on",
"time": "37.808",
"value": "undefined"
},
{
"operation": "hyperlink",
"time": "55.532",
"value": "West Virginia Route 214"
},
{
"operation": "Now on",
"time": "60.587",
"value": "West Virginia Route 214"
},
{
"operation": "Highlighting",
"time": "64.618",
"value": "West Virginia Route 214_sentence_0West Virginia Route 214_sentence_1"
},
{
"operation": "Highlighting",
"time": "74.499",
"value": "West Virginia Route 214_sentence_3"
},
{
"operation": "Highlighting",
"time": "81.664",
"value": "West Virginia Route 214_sentence_4"
},
{
"operation": "finish",
"time": "136.421",
"value": "finish"
}
]
}
|
exfever_8630
|
Interpersonal psychotherapy is a brief, attachment-focused physiotherapy that centers on resolving interpersonal problems and symptomatic recovery developed by an American psychiatrist and researcher.
|
{
"evidence": [
"Interpersonal psychotherapy",
"Gerald Klerman"
],
"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:\nInterpersonal psychotherapy is a brief, attachment-focused physiotherapy that centers on resolving interpersonal problems and symptomatic recovery developed by an American psychiatrist and researcher.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Interpersonal psychotherapy",
"Gerald Klerman"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 4760,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_554
|
Justin Bieber was discovered by an American record executive, talent manager, media proprietor, and businessman.
|
{
"evidence": [
"Justin Bieber",
"Scooter Braun"
],
"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:\nJustin Bieber was discovered by an American record executive, talent manager, media proprietor, and businessman.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Justin Bieber",
"Scooter Braun"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 4761,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
26936
|
Tribal people observe their own festivals and have their own way of life to sever all privileges in matters connected with marriage and succession.
|
{
"evidence": [
"Tribal religions in India_sentence_3",
"Tribal religions in India_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:\nTribal people observe their own festivals and have their own way of life to sever all privileges in matters connected with marriage and succession.\n",
"role": "user"
}
] |
Search terms not in claim
|
{
"ground_truth": {
"evidence": [
"Tribal religions in India_sentence_3",
"Tribal religions in India_sentence_4"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 4762,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.992",
"value": "?search="
},
{
"operation": "search",
"time": "65.542",
"value": "Contains...Tribal people observe their own festivals and have their own way of life"
},
{
"operation": "Now on",
"time": "69.466",
"value": "undefined"
},
{
"operation": "search",
"time": "79.991",
"value": "Contains...observe their own festivals and have their own way of life"
},
{
"operation": "Now on",
"time": "82.121",
"value": "undefined"
},
{
"operation": "search",
"time": "177.316",
"value": "Tribal religions in India"
},
{
"operation": "Now on",
"time": "178.536",
"value": "Tribal religions in India"
},
{
"operation": "Highlighting",
"time": "195.87",
"value": "Tribal religions in India_sentence_3"
},
{
"operation": "Highlighting",
"time": "208.266",
"value": "Tribal religions in India_sentence_4"
},
{
"operation": "finish",
"time": "275.569",
"value": "finish"
}
]
}
|
exfever_220
|
Zindagi Na Milegi Dobara was produced by Farhan Akhtar and a Korean film producer.
|
{
"evidence": [
"Zindagi Na Milegi Dobara",
"Ritesh Sidhwani"
],
"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:\nZindagi Na Milegi Dobara was produced by Farhan Akhtar and a Korean film producer.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Zindagi Na Milegi Dobara",
"Ritesh Sidhwani"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 4763,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_2719
|
The Lesser Sunda Islands are also protected as Komodo National Park.
|
{
"evidence": [
"Bali",
"Lesser Sunda Islands",
"Australia"
],
"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 Lesser Sunda Islands are also protected as Komodo National Park.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Bali",
"Lesser Sunda Islands",
"Australia"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 4764,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_5086
|
Pearl Jam is an American girl band formed in a seaport city on the West Coast of the United States and the seat of King County, Washington.
|
{
"evidence": [
"Pearl Jam",
"Seattle"
],
"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:\nPearl Jam is an American girl band formed in a seaport city on the West Coast of the United States and the seat of King County, Washington.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Pearl Jam",
"Seattle"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 4765,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
548
|
In the 2002 Harrogate Borough Council election, the Liberal Democrats had a minority on the council with 19 councillors only and 19 seats being contested for most of the time in which the changes increased the number of councillors by 20 from 19 to 39.
|
{
"evidence": [
"2002 Harrogate Borough Council election_sentence_3",
"2002 Harrogate Borough Council election_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 2002 Harrogate Borough Council election, the Liberal Democrats had a minority on the council with 19 councillors only and 19 seats being contested for most of the time in which the changes increased the number of councillors by 20 from 19 to 39.\n",
"role": "user"
}
] |
Numerical Reasoning
|
{
"ground_truth": {
"evidence": [
"2002 Harrogate Borough Council election_sentence_3",
"2002 Harrogate Borough Council election_sentence_5"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 4766,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "search",
"time": "15.459",
"value": "2002 Harrogate Borough Council election"
},
{
"operation": "Highlighting",
"time": "118.651",
"value": "2002 Harrogate Borough Council election_sentence_2"
},
{
"operation": "Highlighting",
"time": "216.168",
"value": "2002 Harrogate Borough Council election_sentence_3"
},
{
"operation": "Highlighting deleted",
"time": "231.727",
"value": "2002 Harrogate Borough Council election_sentence_2"
},
{
"operation": "Highlighting",
"time": "256.242",
"value": "2002 Harrogate Borough Council election_sentence_5"
},
{
"operation": "finish",
"time": "263.96",
"value": "finish"
}
]
}
|
75084
|
Pazhou Pagoda, an octagonal tower that covers a total area of 111 square meters, is an ancient Chinese pagoda which is also known as the Whampao Pagoda or Pa Chow Pagoda.
|
{
"evidence": [
"Pazhou Pagoda_sentence_0",
"Pazhou Pagoda_sentence_4",
"Pazhou Pagoda_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:\nPazhou Pagoda, an octagonal tower that covers a total area of 111 square meters, is an ancient Chinese pagoda which is also known as the Whampao Pagoda or Pa Chow Pagoda.\n",
"role": "user"
}
] |
Multi-hop Reasoning
|
{
"ground_truth": {
"evidence": [
"Pazhou Pagoda_sentence_0",
"Pazhou Pagoda_sentence_4",
"Pazhou Pagoda_sentence_6"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 4767,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.111",
"value": "?search="
},
{
"operation": "search",
"time": "144.79",
"value": "Walkden South in Salford, England"
},
{
"operation": "Now on",
"time": "146.56",
"value": "undefined"
},
{
"operation": "search",
"time": "182.777",
"value": "Pazhou Pagoda"
},
{
"operation": "Now on",
"time": "184.002",
"value": "Pazhou Pagoda"
},
{
"operation": "Highlighting",
"time": "209.691",
"value": "Pazhou Pagoda_sentence_0"
},
{
"operation": "Highlighting",
"time": "229.861",
"value": "Pazhou Pagoda_sentence_4"
},
{
"operation": "Highlighting",
"time": "253.855",
"value": "Pazhou Pagoda_sentence_6"
},
{
"operation": "finish",
"time": "306.114",
"value": "finish"
}
]
}
|
exfever_2239
|
The Escape from Planet Earth film was produced by a Canadian computer animation and design firm best known for creating the majority of the film series for a fashion doll manufactured by the American toy company Mattel, Inc. and launched in March 1959.
|
{
"evidence": [
"Escape from Planet Earth",
"Rainmaker Entertainment",
"Barbie"
],
"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 Escape from Planet Earth film was produced by a Canadian computer animation and design firm best known for creating the majority of the film series for a fashion doll manufactured by the American toy company Mattel, Inc. and launched in March 1959.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Escape from Planet Earth",
"Rainmaker Entertainment",
"Barbie"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 4768,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_4436
|
Alfred Hitchcock Presents is the first anthology series that aired on CBS and NBC between 1955 and 1965.
|
{
"evidence": [
"Alfred Hitchcock",
"Alfred Hitchcock Presents",
"Anthology series"
],
"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:\nAlfred Hitchcock Presents is the first anthology series that aired on CBS and NBC between 1955 and 1965.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Alfred Hitchcock",
"Alfred Hitchcock Presents",
"Anthology series"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 4769,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_7389
|
Escape from Planet Earth is the second 3D computer-animated?film produced by Rainmaker Entertainment.
|
{
"evidence": [
"Escape from Planet Earth",
"Comedy film",
"Burlesque"
],
"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:\nEscape from Planet Earth is the second 3D computer-animated?film produced by Rainmaker Entertainment.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Escape from Planet Earth",
"Comedy film",
"Burlesque"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 4770,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_2047
|
Lonesome Dove is an American epic Western adventure television miniseries based on a four-part adaptation of the 1981 novel by Larry McMurtry directed by an American novelist, essayist, bookseller, and screenwriter whose work was predominantly set in either the Old West or contemporary Texas.
|
{
"evidence": [
"Lonesome Dove (miniseries)",
"Larry McMurtry"
],
"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:\nLonesome Dove is an American epic Western adventure television miniseries based on a four-part adaptation of the 1981 novel by Larry McMurtry directed by an American novelist, essayist, bookseller, and screenwriter whose work was predominantly set in either the Old West or contemporary Texas.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Lonesome Dove (miniseries)",
"Larry McMurtry"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 4771,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_2579
|
Sammy Cahn and his collaborators had a series of hit recordings during the singer's tenure at Capitol Records, with an Australian singer and actor, one of the best-selling music artists of all time.
|
{
"evidence": [
"Sammy Cahn",
"Frank Sinatra"
],
"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:\nSammy Cahn and his collaborators had a series of hit recordings during the singer's tenure at Capitol Records, with an Australian singer and actor, one of the best-selling music artists of all time.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Sammy Cahn",
"Frank Sinatra"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 4772,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
53230
|
Aedesius, the father of Margaret "the Virgin," disowned her after embracing Christianity, while her mother died after childbirth.
|
{
"evidence": [
"Aedesius_sentence_1",
"Aedesius of Alexandria_sentence_1",
"Aedesius of Alexandria_sentence_4",
"Aedesius of Alexandria_sentence_3"
],
"label": "NOT ENOUGH INFO"
}
|
feverous
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nAedesius, the father of Margaret \"the Virgin,\" disowned her after embracing Christianity, while her mother died after childbirth.\n",
"role": "user"
}
] |
Entity Disambiguation
|
{
"ground_truth": {
"evidence": [
"Aedesius_sentence_1",
"Aedesius of Alexandria_sentence_1",
"Aedesius of Alexandria_sentence_4",
"Aedesius of Alexandria_sentence_3"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 4773,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.996",
"value": "?search="
},
{
"operation": "search",
"time": "14.117",
"value": "Aedesius"
},
{
"operation": "Now on",
"time": "15.001",
"value": "Aedesius"
},
{
"operation": "Highlighting",
"time": "20.281",
"value": "Aedesius_sentence_1"
},
{
"operation": "search",
"time": "67.878",
"value": "Aedesius of Alexandria"
},
{
"operation": "Now on",
"time": "70.379",
"value": "Aedesius of Alexandria"
},
{
"operation": "Highlighting",
"time": "80.161",
"value": "Aedesius of Alexandria_sentence_1"
},
{
"operation": "Highlighting",
"time": "129.407",
"value": "Aedesius of Alexandria_sentence_7"
},
{
"operation": "Highlighting",
"time": "145.087",
"value": "Aedesius of Alexandria_sentence_4"
},
{
"operation": "Highlighting",
"time": "149.631",
"value": "Aedesius of Alexandria_sentence_3"
},
{
"operation": "Highlighting deleted",
"time": "153.577",
"value": "Aedesius of Alexandria_sentence_3"
},
{
"operation": "Highlighting",
"time": "155.759",
"value": "Aedesius of Alexandria_sentence_3"
},
{
"operation": "Highlighting deleted",
"time": "160.329",
"value": "Aedesius of Alexandria_sentence_7"
},
{
"operation": "finish",
"time": "176.307",
"value": "finish"
}
]
}
|
50553
|
Gyula Ferdinandy, born June 1, 1900, was the Minister of the Interior of Hungary( a member of the Hungarian cabinet and the head of the Ministry of Interior).
|
{
"evidence": [
"Gyula Ferdinandy_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:\nGyula Ferdinandy, born June 1, 1900, was the Minister of the Interior of Hungary( a member of the Hungarian cabinet and the head of the Ministry of Interior).\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Gyula Ferdinandy_sentence_2"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 4774,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.701",
"value": "?search="
},
{
"operation": "search",
"time": "8.777",
"value": "Gyula Ferdinandy"
},
{
"operation": "Now on",
"time": "9.454",
"value": "Gyula Ferdinandy"
},
{
"operation": "Highlighting",
"time": "23.161",
"value": "Gyula Ferdinandy_sentence_2"
},
{
"operation": "finish",
"time": "33.385",
"value": "finish"
}
]
}
|
54798
|
Carl Sully graduated from Macquarie University(a public research university based in Sydney, Australia), and was a member of the New South Wales Parliament for Smithfield.
|
{
"evidence": [
"Macquarie University_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:\nCarl Sully graduated from Macquarie University(a public research university based in Sydney, Australia), and was a member of the New South Wales Parliament for Smithfield.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Macquarie University_sentence_0"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 4775,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "2.3",
"value": "?search="
},
{
"operation": "search",
"time": "61.108",
"value": "Carl Sullivan"
},
{
"operation": "Now on",
"time": "62.83",
"value": "Carl Sullivan"
},
{
"operation": "search",
"time": "85.151",
"value": "Contains...carl sully"
},
{
"operation": "Now on",
"time": "86.687",
"value": "undefined"
},
{
"operation": "search",
"time": "93.78",
"value": "\"Carl Sully\""
},
{
"operation": "Now on",
"time": "102.234",
"value": "undefined"
},
{
"operation": "search",
"time": "181.099",
"value": "New South Wales Parliament"
},
{
"operation": "Now on",
"time": "182.686",
"value": "Parliament of New South Wales"
},
{
"operation": "Page search",
"time": "194.91",
"value": "Sully"
},
{
"operation": "hyperlink",
"time": "219.929",
"value": "Parliaments of the Australian states and territories"
},
{
"operation": "Now on",
"time": "221.222",
"value": "Parliaments of the Australian states and territories"
},
{
"operation": "hyperlink",
"time": "242.174",
"value": "2.1 New South Wales Parliaments of the Australian states and territories_section_2"
},
{
"operation": "search",
"time": "294.881",
"value": "Macquarie University"
},
{
"operation": "Now on",
"time": "296.829",
"value": "Macquarie University"
},
{
"operation": "Highlighting",
"time": "300.784",
"value": "Macquarie University_sentence_0"
},
{
"operation": "search",
"time": "351.602",
"value": "\"Carl Sully\""
},
{
"operation": "Now on",
"time": "352.695",
"value": "undefined"
},
{
"operation": "finish",
"time": "371.127",
"value": "finish"
}
]
}
|
54798
|
Carl Sully graduated from Macquarie University(a public research university based in Sydney, Australia), and was a member of the New South Wales Parliament for Smithfield.
|
{
"evidence": [
"Macquarie University_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:\nCarl Sully graduated from Macquarie University(a public research university based in Sydney, Australia), and was a member of the New South Wales Parliament for Smithfield.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Macquarie University_sentence_0"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 4776,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "2.3",
"value": "?search="
},
{
"operation": "search",
"time": "61.108",
"value": "Carl Sullivan"
},
{
"operation": "Now on",
"time": "62.83",
"value": "Carl Sullivan"
},
{
"operation": "search",
"time": "85.151",
"value": "Contains...carl sully"
},
{
"operation": "Now on",
"time": "86.687",
"value": "undefined"
},
{
"operation": "search",
"time": "93.78",
"value": "\"Carl Sully\""
},
{
"operation": "Now on",
"time": "102.234",
"value": "undefined"
},
{
"operation": "search",
"time": "181.099",
"value": "New South Wales Parliament"
},
{
"operation": "Now on",
"time": "182.686",
"value": "Parliament of New South Wales"
},
{
"operation": "Page search",
"time": "194.91",
"value": "Sully"
},
{
"operation": "hyperlink",
"time": "219.929",
"value": "Parliaments of the Australian states and territories"
},
{
"operation": "Now on",
"time": "221.222",
"value": "Parliaments of the Australian states and territories"
},
{
"operation": "hyperlink",
"time": "242.174",
"value": "2.1 New South Wales Parliaments of the Australian states and territories_section_2"
},
{
"operation": "search",
"time": "294.881",
"value": "Macquarie University"
},
{
"operation": "Now on",
"time": "296.829",
"value": "Macquarie University"
},
{
"operation": "Highlighting",
"time": "300.784",
"value": "Macquarie University_sentence_0"
},
{
"operation": "search",
"time": "351.602",
"value": "\"Carl Sully\""
},
{
"operation": "Now on",
"time": "352.695",
"value": "undefined"
},
{
"operation": "finish",
"time": "371.127",
"value": "finish"
}
]
}
|
12948
|
Salvatore Borgh signed with Hamilton Thunder (a soccer team located in Hamilton, Ontario, playing in the Western Conference of the Canadian Professional Soccer League) in 2002 but was loaned to Calgary Storm until July 21, 2002.
|
{
"evidence": [
"Salvatore Borgh_sentence_7",
"Hamilton Thunder_sentence_0",
"Salvatore Borgh_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:\nSalvatore Borgh signed with Hamilton Thunder (a soccer team located in Hamilton, Ontario, playing in the Western Conference of the Canadian Professional Soccer League) in 2002 but was loaned to Calgary Storm until July 21, 2002.\n",
"role": "user"
}
] |
Multi-hop Reasoning
|
{
"ground_truth": {
"evidence": [
"Salvatore Borgh_sentence_7",
"Hamilton Thunder_sentence_0",
"Salvatore Borgh_sentence_9"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 4777,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.922",
"value": "?search="
},
{
"operation": "search",
"time": "10.247",
"value": "Salvatore Borgh"
},
{
"operation": "Now on",
"time": "11.092",
"value": "Salvatore Borgh"
},
{
"operation": "Page search",
"time": "51.138",
"value": "hamilton thunder"
},
{
"operation": "Highlighting",
"time": "53.801",
"value": "Salvatore Borgh_sentence_7"
},
{
"operation": "hyperlink",
"time": "62.11",
"value": "Hamilton Thunder"
},
{
"operation": "Now on",
"time": "62.558",
"value": "Hamilton Thunder"
},
{
"operation": "Highlighting",
"time": "64.115",
"value": "Hamilton Thunder_sentence_0"
},
{
"operation": "back-button-clicked",
"time": "70.909",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "70.979",
"value": "Salvatore Borgh"
},
{
"operation": "Highlighting",
"time": "143.407",
"value": "Salvatore Borgh_sentence_9"
},
{
"operation": "finish",
"time": "154.509",
"value": "finish"
}
]
}
|
exfever_5370
|
Neil Armstrong was the first project pilot on the Century Series fighters.
|
{
"evidence": [
"Neil Armstrong",
"Astronaut"
],
"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:\nNeil Armstrong was the first project pilot on the Century Series fighters.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Neil Armstrong",
"Astronaut"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 4778,
"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": 4779,
"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"
}
]
}
|
exfever_8001
|
Boromir is a fictional character in J. R. R. Tolkien's legendarium and the elder brother of the second son of the 27th Ruling Steward of Gondor.
|
{
"evidence": [
"Boromir",
"Faramir",
"Denethor"
],
"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:\nBoromir is a fictional character in J. R. R. Tolkien's legendarium and the elder brother of the second son of the 27th Ruling Steward of Gondor.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Boromir",
"Faramir",
"Denethor"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 4780,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_5008
|
Chris Hardwick is the first American comedian to host Talking Dead, an hourlong aftershow on AMC.
|
{
"evidence": [
"Chris Hardwick",
"Fear the Walking Dead",
"AMC (TV channel)"
],
"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:\nChris Hardwick is the first American comedian to host Talking Dead, an hourlong aftershow on AMC.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Chris Hardwick",
"Fear the Walking Dead",
"AMC (TV channel)"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 4781,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_8619
|
Same Old Love is a song by an actress who starred in Wizards of Waverly Place as a fictional character and the protagonist of the Disney Channel sitcom, Wizards of Waverly Place.
|
{
"evidence": [
"Same Old Love",
"Selena Gomez",
"Alex Russo"
],
"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:\nSame Old Love is a song by an actress who starred in Wizards of Waverly Place as a fictional character and the protagonist of the Disney Channel sitcom, Wizards of Waverly Place.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Same Old Love",
"Selena Gomez",
"Alex Russo"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 4782,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_5904
|
The Magnificent Ambersons, directed by Orson Welles, received an Academy Award.
|
{
"evidence": [
"Orson Welles",
"The Mercury Theatre on the Air"
],
"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 Magnificent Ambersons, directed by Orson Welles, received an Academy Award.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Orson Welles",
"The Mercury Theatre on the Air"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 4783,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.