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 |
|---|---|---|---|---|---|---|---|---|
59503
|
Beckworth was elected as a Democrat to the 76th Congress, but was not a candidate for renomination in 1952; and was never elected to the 85th Congress.
|
{
"evidence": [
"Lindley Beckworth_sentence_9"
],
"label": "REFUTE"
}
|
feverous
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nBeckworth was elected as a Democrat to the 76th Congress, but was not a candidate for renomination in 1952; and was never elected to the 85th Congress.\n",
"role": "user"
}
] |
Entity Disambiguation
|
{
"ground_truth": {
"evidence": [
"Lindley Beckworth_sentence_9"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 7784,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.091",
"value": "?search="
},
{
"operation": "search",
"time": "8.354",
"value": "Beckworth"
},
{
"operation": "Now on",
"time": "9.14",
"value": "Beckworth"
},
{
"operation": "hyperlink",
"time": "14.191",
"value": "Lindley Beckworth"
},
{
"operation": "Now on",
"time": "14.97",
"value": "Lindley Beckworth"
},
{
"operation": "Page search",
"time": "19.092",
"value": "76"
},
{
"operation": "Highlighting",
"time": "25.461",
"value": "Lindley Beckworth_sentence_7"
},
{
"operation": "Highlighting",
"time": "39.278",
"value": "Lindley Beckworth_sentence_8"
},
{
"operation": "Highlighting deleted",
"time": "45.997",
"value": "Lindley Beckworth_sentence_8"
},
{
"operation": "Highlighting deleted",
"time": "47.162",
"value": "Lindley Beckworth_sentence_7"
},
{
"operation": "Highlighting",
"time": "48.299",
"value": "Lindley Beckworth_sentence_9"
},
{
"operation": "finish",
"time": "199.846",
"value": "finish"
}
]
}
|
2898
|
There were two churches that got closed in 2013, the first one was St. Andrew in Colwyn Bay, which was founded in 1908, and St. Catherine in Old Colwyn, which was founded in 1837.
|
{
"evidence": [
"St. Andrew's Church_sentence_0",
"St. Catherine's Church_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:\nThere were two churches that got closed in 2013, the first one was St. Andrew in Colwyn Bay, which was founded in 1908, and St. Catherine in Old Colwyn, which was founded in 1837.\n",
"role": "user"
}
] |
Multi-hop Reasoning
|
{
"ground_truth": {
"evidence": [
"St. Andrew's Church_sentence_0",
"St. Catherine's Church_sentence_0"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 7785,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "search",
"time": "24.517",
"value": "st. andrew colwyn bay"
},
{
"operation": "search",
"time": "37.766",
"value": "St. Andrew's Church"
},
{
"operation": "Page search",
"time": "47.838",
"value": "closed"
},
{
"operation": "Page search",
"time": "60.218",
"value": "colwyn bay"
},
{
"operation": "search",
"time": "93.641",
"value": "Colwyn Bay"
},
{
"operation": "Page search",
"time": "104.521",
"value": "st andrew"
},
{
"operation": "search",
"time": "118.827",
"value": "church closures 2013"
},
{
"operation": "hyperlink",
"time": "122.383",
"value": "church closure 2013"
},
{
"operation": "Page search",
"time": "130.15",
"value": "st. andrew"
},
{
"operation": "search",
"time": "146.38",
"value": "old colwyn"
},
{
"operation": "hyperlink",
"time": "173.19",
"value": "2 History Old Colwyn_section_1"
},
{
"operation": "Page search",
"time": "181.367",
"value": "2013"
},
{
"operation": "search",
"time": "191.315",
"value": "church closure"
},
{
"operation": "Page search",
"time": "197.362",
"value": "2013"
},
{
"operation": "hyperlink",
"time": "204.775",
"value": "Church closure Alan Campbell (pastor)_section_1"
},
{
"operation": "Highlighting",
"time": "210.668",
"value": "Alan Campbell (pastor)_sentence_18"
},
{
"operation": "Highlighting",
"time": "210.672",
"value": "Alan Campbell (pastor)_title"
},
{
"operation": "Highlighting",
"time": "210.674",
"value": "Alan Campbell (pastor)_section_1"
},
{
"operation": "back-button-clicked",
"time": "236.222",
"value": "back-button-clicked"
},
{
"operation": "hyperlink",
"time": "245.707",
"value": "St Mary's Church, Rogiet"
},
{
"operation": "Highlighting",
"time": "250.463",
"value": "St Mary's Church, Rogiet_sentence_10"
},
{
"operation": "Highlighting",
"time": "250.467",
"value": "St Mary's Church, Rogiet_title"
},
{
"operation": "Highlighting",
"time": "250.47",
"value": "St Mary's Church, Rogiet_section_0"
},
{
"operation": "back-button-clicked",
"time": "252.818",
"value": "back-button-clicked"
},
{
"operation": "hyperlink",
"time": "258.894",
"value": "St. George's (Round) Church, Halifax, Nova Scotia"
},
{
"operation": "Page search",
"time": "271.097",
"value": "closure"
},
{
"operation": "back-button-clicked",
"time": "281.019",
"value": "back-button-clicked"
},
{
"operation": "Page search",
"time": "287.947",
"value": "2013"
},
{
"operation": "hyperlink",
"time": "290.589",
"value": "Holy Trinity Church, Brighton"
},
{
"operation": "back-button-clicked",
"time": "305.705",
"value": "back-button-clicked"
},
{
"operation": "search",
"time": "327.866",
"value": "St. Andrew's Church"
},
{
"operation": "Page search",
"time": "333.485",
"value": "closure"
},
{
"operation": "hyperlink",
"time": "357.007",
"value": "25.4 Wales St. Andrew's Church_section_63"
},
{
"operation": "Highlighting",
"time": "371.196",
"value": "St. Andrew's Church_sentence_0"
},
{
"operation": "search",
"time": "402.904",
"value": "St. Catherine Church, Birmingham"
},
{
"operation": "back-button-clicked",
"time": "416.374",
"value": "back-button-clicked"
},
{
"operation": "Page search",
"time": "423.595",
"value": "1908"
},
{
"operation": "search",
"time": "444.417",
"value": "st. catherine's church"
},
{
"operation": "Highlighting",
"time": "450.114",
"value": "St. Catherine's Church_sentence_0"
},
{
"operation": "finish",
"time": "483.321",
"value": "finish"
}
]
}
|
89198
|
I Am a Hero(April 27, 2009 – February 27, 2017) is a Japanese vampire manga series written and illustrated by Kengo Hanazawa and published by Shogakukan.
|
{
"evidence": [
"I Am a Hero_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:\nI Am a Hero(April 27, 2009 – February 27, 2017) is a Japanese vampire manga series written and illustrated by Kengo Hanazawa and published by Shogakukan.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"I Am a Hero_sentence_0"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 7786,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.203",
"value": "?search="
},
{
"operation": "search",
"time": "11.986",
"value": "I Am a Hero"
},
{
"operation": "Now on",
"time": "13.557",
"value": "I Am a Hero"
},
{
"operation": "Highlighting",
"time": "53.569",
"value": "I Am a Hero_sentence_0"
},
{
"operation": "Page search",
"time": "59.804",
"value": "vamp"
},
{
"operation": "finish",
"time": "67.601",
"value": "finish"
}
]
}
|
8134
|
Wilfred Brimble was an only child born in California.
|
{
"evidence": [
"Wilfred Brimble_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:\nWilfred Brimble was an only child born in California.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Wilfred Brimble_sentence_4"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 7787,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.57",
"value": "?search="
},
{
"operation": "search",
"time": "27.86",
"value": "wilfred brimble"
},
{
"operation": "Now on",
"time": "28.818",
"value": "Wilfred Brimble"
},
{
"operation": "search",
"time": "59.683",
"value": "brimble"
},
{
"operation": "Now on",
"time": "60.411",
"value": "Brimble"
},
{
"operation": "hyperlink",
"time": "73.266",
"value": "Wilfred Brimble"
},
{
"operation": "Now on",
"time": "73.652",
"value": "Wilfred Brimble"
},
{
"operation": "Highlighting",
"time": "79.567",
"value": "Wilfred Brimble_sentence_7"
},
{
"operation": "Highlighting",
"time": "79.573",
"value": "Wilfred Brimble_section_1"
},
{
"operation": "Highlighting",
"time": "93.243",
"value": "Wilfred Brimble_sentence_4"
},
{
"operation": "Highlighting",
"time": "111.169",
"value": "Wilfred Brimble_sentence_4"
},
{
"operation": "finish",
"time": "116.076",
"value": "finish"
}
]
}
|
exfever_5554
|
David Thewlis's other film roles include a 1995 erotic historic drama film based on a 1967 play by a Portuguese-born American playwright, screenwriter, translator, and film director.
|
{
"evidence": [
"David Thewlis",
"Total Eclipse (film)",
"Christopher Hampton"
],
"label": "REFUTE"
}
|
exfever
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nDavid Thewlis's other film roles include a 1995 erotic historic drama film based on a 1967 play by a Portuguese-born American playwright, screenwriter, translator, and film director.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"David Thewlis",
"Total Eclipse (film)",
"Christopher Hampton"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 7788,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
6396
|
"Body and Soul", sung by Kennedy (an Australian actress, singer, film director and screenwriter) and written as "She Got Body She Got Soul" by Tim Finn of Split Enz, made number five on the Australian Kent Music Report Singles Chart in May 1982
|
{
"evidence": [
"Jo Kennedy_sentence_1",
"Jo Kennedy_sentence_4",
"Jo Kennedy_sentence_5",
"Jo Kennedy_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:\n\"Body and Soul\", sung by Kennedy (an Australian actress, singer, film director and screenwriter) and written as \"She Got Body She Got Soul\" by Tim Finn of Split Enz, made number five on the Australian Kent Music Report Singles Chart in May 1982\n",
"role": "user"
}
] |
Entity Disambiguation
|
{
"ground_truth": {
"evidence": [
"Jo Kennedy_sentence_1",
"Jo Kennedy_sentence_4",
"Jo Kennedy_sentence_5",
"Jo Kennedy_sentence_6"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 7789,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.832",
"value": "?search="
},
{
"operation": "search",
"time": "21.916",
"value": "Kennedy"
},
{
"operation": "Now on",
"time": "23.453",
"value": "Kennedy"
},
{
"operation": "hyperlink",
"time": "33.049",
"value": "Kennedy (commentator)"
},
{
"operation": "Now on",
"time": "34.251",
"value": "Kennedy (commentator)"
},
{
"operation": "Highlighting",
"time": "39.527",
"value": "Kennedy (commentator)_sentence_0"
},
{
"operation": "Highlighting",
"time": "39.554",
"value": "Kennedy (commentator)_title"
},
{
"operation": "Page search",
"time": "63.969",
"value": "body and soul"
},
{
"operation": "back-button-clicked",
"time": "72.641",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "72.965",
"value": "Kennedy"
},
{
"operation": "search",
"time": "112.277",
"value": "Jo Kennedy"
},
{
"operation": "Now on",
"time": "113.54",
"value": "Jo Kennedy"
},
{
"operation": "Highlighting",
"time": "146.384",
"value": "Jo Kennedy_sentence_1"
},
{
"operation": "Highlighting",
"time": "159.786",
"value": "Jo Kennedy_header_cell_1_2_1"
},
{
"operation": "Highlighting",
"time": "159.81",
"value": "Jo Kennedy_section_0"
},
{
"operation": "Highlighting",
"time": "159.823",
"value": "Jo Kennedy_section_1"
},
{
"operation": "Highlighting",
"time": "170.046",
"value": "Jo Kennedy_sentence_4"
},
{
"operation": "Highlighting",
"time": "171.719",
"value": "Jo Kennedy_sentence_5"
},
{
"operation": "Highlighting",
"time": "175.264",
"value": "Jo Kennedy_sentence_6"
},
{
"operation": "Highlighting deleted",
"time": "178.926",
"value": "Jo Kennedy_header_cell_1_2_1"
},
{
"operation": "Highlighting deleted",
"time": "178.97",
"value": "Jo Kennedy_section_0"
},
{
"operation": "Highlighting deleted",
"time": "178.998",
"value": "Jo Kennedy_section_1"
},
{
"operation": "finish",
"time": "183.068",
"value": "finish"
}
]
}
|
exfever_3532
|
Prison Break is an American television series co-produced by an American director and actor who is notable for his role in a 1975 film written by Gerald Sanford.
|
{
"evidence": [
"Prison Break",
"Kevin Hooks",
"Aaron Loves Angela"
],
"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:\nPrison Break is an American television series co-produced by an American director and actor who is notable for his role in a 1975 film written by Gerald Sanford.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Prison Break",
"Kevin Hooks",
"Aaron Loves Angela"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 7790,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_7847
|
Jane Austen's Emma is an adaptation of the 1815 novel of the same name, starring an English actress and model who has appeared in British costume dramas, including a 2000 period romantic drama film directed by Mel Gibson.
|
{
"evidence": [
"Emma (1996 TV film)",
"Kate Beckinsale",
"The Golden Bowl (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:\nJane Austen's Emma is an adaptation of the 1815 novel of the same name, starring an English actress and model who has appeared in British costume dramas, including a 2000 period romantic drama film directed by Mel Gibson.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Emma (1996 TV film)",
"Kate Beckinsale",
"The Golden Bowl (film)"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 7791,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_8206
|
Nymphomaniac was a high-grossing film directed by Lars von Trier.
|
{
"evidence": [
"Nymphomaniac (film)",
"Uma Thurman",
"Academy Awards"
],
"label": "NOT ENOUGH INFO"
}
|
exfever
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nNymphomaniac was a high-grossing film directed by Lars von Trier.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Nymphomaniac (film)",
"Uma Thurman",
"Academy Awards"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 7792,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
20626
|
In August 2004, Nicholas Monroe, who won three ATP World Tour titles, had first reached his first semifinals at Indonesia F2 in Makassar, then in October of the same year, he captured his first doubles Futures title, in doubles, partnering Márcio Torres, at Venezuela F3.
|
{
"evidence": [
"Nicholas Monroe_sentence_1",
"Nicholas Monroe_sentence_11",
"Nicholas Monroe_sentence_12"
],
"label": "SUPPORT"
}
|
feverous
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nIn August 2004, Nicholas Monroe, who won three ATP World Tour titles, had first reached his first semifinals at Indonesia F2 in Makassar, then in October of the same year, he captured his first doubles Futures title, in doubles, partnering Márcio Torres, at Venezuela F3.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Nicholas Monroe_sentence_1",
"Nicholas Monroe_sentence_11",
"Nicholas Monroe_sentence_12"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 7793,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "2.442",
"value": "?search="
},
{
"operation": "search",
"time": "14.545",
"value": "Nicholas Monroe"
},
{
"operation": "Now on",
"time": "19.928",
"value": "Nicholas Monroe"
},
{
"operation": "Highlighting",
"time": "36.798",
"value": "Nicholas Monroe_sentence_1"
},
{
"operation": "Page search",
"time": "47.421",
"value": "2004"
},
{
"operation": "Page search",
"time": "60.494",
"value": "semi"
},
{
"operation": "Highlighting",
"time": "69.546",
"value": "Nicholas Monroe_sentence_11"
},
{
"operation": "Highlighting",
"time": "83.489",
"value": "Nicholas Monroe_sentence_12"
},
{
"operation": "finish",
"time": "97.559",
"value": "finish"
}
]
}
|
exfever_3318
|
Jonah Hill is known for his comedic roles in films such as the 2010 American comedy film written, produced, and directed by Nicholas Stoller.
|
{
"evidence": [
"Jonah Hill",
"Get Him to the Greek"
],
"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:\nJonah Hill is known for his comedic roles in films such as the 2010 American comedy film written, produced, and directed by Nicholas Stoller.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Jonah Hill",
"Get Him to the Greek"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 7794,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
67585
|
Willem Vrolik, a Dutch anatomist and pathologist who is a pioneer in the field of vertebrate teratology, published teratological works on cyclopia, the pathogenesis of congenital anomalies, and a treatise on conjoined twins as well as Handboek der ziektekundige ontleedkunde (Handbook of pathological anatomy); however, these were later on stopped in production.
|
{
"evidence": [
"Willem Vrolik_sentence_0",
"Willem Vrolik_sentence_1",
"Willem Vrolik_sentence_10",
"Willem Vrolik_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:\nWillem Vrolik, a Dutch anatomist and pathologist who is a pioneer in the field of vertebrate teratology, published teratological works on cyclopia, the pathogenesis of congenital anomalies, and a treatise on conjoined twins as well as Handboek der ziektekundige ontleedkunde (Handbook of pathological anatomy); however, these were later on stopped in production.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Willem Vrolik_sentence_0",
"Willem Vrolik_sentence_1",
"Willem Vrolik_sentence_10",
"Willem Vrolik_sentence_11"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 7795,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.061",
"value": "?search="
},
{
"operation": "search",
"time": "8.913",
"value": "Willem Vrolik"
},
{
"operation": "Now on",
"time": "10.331",
"value": "Willem Vrolik"
},
{
"operation": "Highlighting",
"time": "723.861",
"value": "Willem Vrolik_sentence_0"
},
{
"operation": "Highlighting",
"time": "727.076",
"value": "Willem Vrolik_sentence_1"
},
{
"operation": "Highlighting",
"time": "734.464",
"value": "Willem Vrolik_sentence_10"
},
{
"operation": "Highlighting",
"time": "747.617",
"value": "Willem Vrolik_sentence_11"
},
{
"operation": "finish",
"time": "804.963",
"value": "finish"
}
]
}
|
8856
|
Lucas Scott had a failed audition by Chad Michael Murry who felt a connection with the character, what was what the producers planned.
|
{
"evidence": [
"Lucas Scott_sentence_7",
"Lucas Scott_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:\nLucas Scott had a failed audition by Chad Michael Murry who felt a connection with the character, what was what the producers planned.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Lucas Scott_sentence_7",
"Lucas Scott_sentence_12"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 7796,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.588",
"value": "?search="
},
{
"operation": "search",
"time": "18.073",
"value": "lucas scott"
},
{
"operation": "Now on",
"time": "20.717",
"value": "Lucas Scott"
},
{
"operation": "Highlighting",
"time": "85.785",
"value": "Lucas Scott_sentence_7"
},
{
"operation": "Highlighting",
"time": "101.547",
"value": "Lucas Scott_sentence_12"
},
{
"operation": "finish",
"time": "169.178",
"value": "finish"
}
]
}
|
exfever_1367
|
Raimi's Spider-Man series won various awards.
|
{
"evidence": [
"Spider-Man 2",
"Spider-Man (2002 film)"
],
"label": "NOT ENOUGH INFO"
}
|
exfever
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nRaimi's Spider-Man series won various awards.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Spider-Man 2",
"Spider-Man (2002 film)"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 7797,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_4594
|
Zayn Malik is a British singer who adopts more alternative R&B music style on his debut studio album, released on 25 March 2016 by RCA Records.
|
{
"evidence": [
"Zayn Malik",
"Mind of Mine"
],
"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:\nZayn Malik is a British singer who adopts more alternative R&B music style on his debut studio album, released on 25 March 2016 by RCA Records.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Zayn Malik",
"Mind of Mine"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 7798,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_3702
|
The Economic Times is an Indian English-language business-focused daily newspaper published simultaneously in 15 cities, including a city and a union territory of India whose city proper population was over 11 million according to the 2011 census.
|
{
"evidence": [
"The Economic Times",
"Delhi"
],
"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 Economic Times is an Indian English-language business-focused daily newspaper published simultaneously in 15 cities, including a city and a union territory of India whose city proper population was over 11 million according to the 2011 census.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"The Economic Times",
"Delhi"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 7799,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_7145
|
Rock music is a body of music which is distinguishable from popular folk music.
|
{
"evidence": [
"Heavy metal music",
"Rock music"
],
"label": "NOT ENOUGH INFO"
}
|
exfever
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nRock music is a body of music which is distinguishable from popular folk music.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Heavy metal music",
"Rock music"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 7800,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_3296
|
Miranda Cosgrove is an American actress and singer who rose to prominence by starring as Carly Shay in an American teen sitcom created by Dan Schneider.
|
{
"evidence": [
"Miranda Cosgrove",
"ICarly"
],
"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:\nMiranda Cosgrove is an American actress and singer who rose to prominence by starring as Carly Shay in an American teen sitcom created by Dan Schneider.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Miranda Cosgrove",
"ICarly"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 7801,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
66588
|
The List of WCW World Television Champions was a professional wrestling world television championship which was sold to the World Wrestling Federation, an American integrated media and entertainment company that is known for professional wrestling.
|
{
"evidence": [
"List of WCW World Television Champions_sentence_0",
"List of WCW World Television Champions_sentence_3",
"WWE_sentence_0"
],
"label": "SUPPORT"
}
|
feverous
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nThe List of WCW World Television Champions was a professional wrestling world television championship which was sold to the World Wrestling Federation, an American integrated media and entertainment company that is known for professional wrestling.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"List of WCW World Television Champions_sentence_0",
"List of WCW World Television Champions_sentence_3",
"WWE_sentence_0"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 7802,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.083",
"value": "?search="
},
{
"operation": "search",
"time": "9.868",
"value": "List of WCW World Television Champions"
},
{
"operation": "Now on",
"time": "11.151",
"value": "List of WCW World Television Champions"
},
{
"operation": "Highlighting",
"time": "13.092",
"value": "List of WCW World Television Champions_sentence_0"
},
{
"operation": "Highlighting",
"time": "24.321",
"value": "List of WCW World Television Champions_sentence_3"
},
{
"operation": "hyperlink",
"time": "31.27",
"value": "World Wrestling Federation"
},
{
"operation": "Now on",
"time": "34.278",
"value": "WWE"
},
{
"operation": "Highlighting",
"time": "35.306",
"value": "WWE_sentence_0"
},
{
"operation": "back-button-clicked",
"time": "39.473",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "39.61",
"value": "List of WCW World Television Champions"
},
{
"operation": "finish",
"time": "64.934",
"value": "finish"
}
]
}
|
exfever_3507
|
Canada is the largest country in North America.
|
{
"evidence": [
"The Secret Life of Us",
"New Zealand"
],
"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:\nCanada is the largest country in North America.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"The Secret Life of Us",
"New Zealand"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 7803,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_2800
|
Bertrand Russell was a British philosopher and social critic who also studied correct reasoning or good arguments, which are often defined in a more narrow sense as the science of deductively valid inferences.
|
{
"evidence": [
"Bertrand Russell",
"Logic"
],
"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:\nBertrand Russell was a British philosopher and social critic who also studied correct reasoning or good arguments, which are often defined in a more narrow sense as the science of deductively valid inferences.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Bertrand Russell",
"Logic"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 7804,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
31122
|
1937–38 AHL season saw Jack Markle top the scoring list for the season with 54 points, followed by Eddie Convey with 52, and Lorne Duguid in third with 49 points.
|
{
"evidence": [
"Jack Markle_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:\n1937–38 AHL season saw Jack Markle top the scoring list for the season with 54 points, followed by Eddie Convey with 52, and Lorne Duguid in third with 49 points.\n",
"role": "user"
}
] |
Search terms not in claim
|
{
"ground_truth": {
"evidence": [
"Jack Markle_sentence_1"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 7805,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.666",
"value": "?search="
},
{
"operation": "search",
"time": "24.227",
"value": "Jack Markle"
},
{
"operation": "Now on",
"time": "25.083",
"value": "Jack Markle"
},
{
"operation": "Highlighting",
"time": "47.669",
"value": "Jack Markle_sentence_1"
},
{
"operation": "search",
"time": "57.083",
"value": "Toronto Maple Leafs"
},
{
"operation": "Now on",
"time": "58.187",
"value": "Toronto Maple Leafs"
},
{
"operation": "search",
"time": "117.194",
"value": "AHL"
},
{
"operation": "Now on",
"time": "118.181",
"value": "AHL"
},
{
"operation": "hyperlink",
"time": "127.52",
"value": "American Hockey League"
},
{
"operation": "Now on",
"time": "128.518",
"value": "American Hockey League"
},
{
"operation": "search",
"time": "214.813",
"value": "Jack Markle"
},
{
"operation": "Now on",
"time": "215.409",
"value": "Jack Markle"
},
{
"operation": "hyperlink",
"time": "240.453",
"value": "National Hockey League"
},
{
"operation": "Now on",
"time": "241.831",
"value": "National Hockey League"
},
{
"operation": "search",
"time": "331.802",
"value": "American Hockey League"
},
{
"operation": "Now on",
"time": "332.564",
"value": "American Hockey League"
},
{
"operation": "finish",
"time": "446.135",
"value": "finish"
}
]
}
|
57224
|
Monica Sims spent three seasons working in theatres in Windsor, she then moved to BBC Television but never returned to radio.
|
{
"evidence": [
"Monica Sims_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:\nMonica Sims spent three seasons working in theatres in Windsor, she then moved to BBC Television but never returned to radio.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Monica Sims_sentence_4"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 7806,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.246",
"value": "?search="
},
{
"operation": "search",
"time": "59.039",
"value": "Monica Sims"
},
{
"operation": "Now on",
"time": "60.008",
"value": "Monica Sims"
},
{
"operation": "Highlighting",
"time": "66.054",
"value": "Monica Sims_sentence_2"
},
{
"operation": "Highlighting",
"time": "77.405",
"value": "Monica Sims_sentence_4"
},
{
"operation": "Highlighting deleted",
"time": "79.001",
"value": "Monica Sims_sentence_2"
},
{
"operation": "finish",
"time": "96.472",
"value": "finish"
}
]
}
|
exfever_2861
|
Miyu Irino is a Japanese voice actor and singer who voiced a fictional character in the manga and anime series Eyeshield 21, whose running ability improves, which was noticed by a fictional character created by Riichiro Inagaki and Yusuke Murata.
|
{
"evidence": [
"Miyu Irino",
"Sena Kobayakawa",
"Yoichi Hiruma"
],
"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:\nMiyu Irino is a Japanese voice actor and singer who voiced a fictional character in the manga and anime series Eyeshield 21, whose running ability improves, which was noticed by a fictional character created by Riichiro Inagaki and Yusuke Murata.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Miyu Irino",
"Sena Kobayakawa",
"Yoichi Hiruma"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 7807,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
47080
|
Deutsche Schule Santiago has two campuses located in Vitacura and Las Condes.
|
{
"evidence": [
"Deutsche Schule Santiago_sentence_1",
"Deutsche Schule Santiago_sentence_2",
"Deutsche Schule Santiago_sentence_3",
"Las Condes_sentence_16"
],
"label": "REFUTE"
}
|
feverous
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nDeutsche Schule Santiago has two campuses located in Vitacura and Las Condes.\n",
"role": "user"
}
] |
Multi-hop Reasoning
|
{
"ground_truth": {
"evidence": [
"Deutsche Schule Santiago_sentence_1",
"Deutsche Schule Santiago_sentence_2",
"Deutsche Schule Santiago_sentence_3",
"Las Condes_sentence_16"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 7808,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.24",
"value": "?search="
},
{
"operation": "search",
"time": "17.656",
"value": "Deutsche Schule Santiago"
},
{
"operation": "Now on",
"time": "19.425",
"value": "Deutsche Schule Santiago"
},
{
"operation": "Highlighting",
"time": "39.856",
"value": "Deutsche Schule Santiago_sentence_1"
},
{
"operation": "Highlighting",
"time": "45.615",
"value": "Deutsche Schule Santiago_sentence_2"
},
{
"operation": "Highlighting",
"time": "47.486",
"value": "Deutsche Schule Santiago_sentence_3"
},
{
"operation": "hyperlink",
"time": "82.847",
"value": "Las Condes"
},
{
"operation": "Now on",
"time": "84.25",
"value": "Las Condes"
},
{
"operation": "hyperlink",
"time": "90.186",
"value": "5 Education Las Condes_section_5"
},
{
"operation": "Highlighting",
"time": "96.287",
"value": "Las Condes_sentence_16"
},
{
"operation": "finish",
"time": "127.946",
"value": "finish"
}
]
}
|
exfever_8996
|
Andrew Jackson was born in the colonial Carolinas in the decade before the American Revolutionary War, became a frontier lawyer, and married the United States' 7th President, who died just days after his election and before his inauguration in 1828.
|
{
"evidence": [
"Andrew Jackson",
"Rachel Jackson"
],
"label": "REFUTE"
}
|
exfever
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nAndrew Jackson was born in the colonial Carolinas in the decade before the American Revolutionary War, became a frontier lawyer, and married the United States' 7th President, who died just days after his election and before his inauguration in 1828.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Andrew Jackson",
"Rachel Jackson"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 7809,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
94070
|
Nebraska Cornhuskers won four times and lost three times during the games they participated in.
|
{
"evidence": [
"Nebraska Cornhuskers_sentence_2"
],
"label": "NOT ENOUGH INFO"
}
|
feverous
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nNebraska Cornhuskers won four times and lost three times during the games they participated in.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Nebraska Cornhuskers_sentence_2"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 7810,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.115",
"value": "?search="
},
{
"operation": "search",
"time": "6.912",
"value": "Nebraska Cornhuskers"
},
{
"operation": "Now on",
"time": "9.041",
"value": "Nebraska Cornhuskers"
},
{
"operation": "Page search",
"time": "18.109",
"value": "won"
},
{
"operation": "Page search",
"time": "35.924",
"value": "won4-3"
},
{
"operation": "Page search",
"time": "39.298",
"value": "4-3"
},
{
"operation": "Page search",
"time": "68.789",
"value": "participate"
},
{
"operation": "Highlighting",
"time": "77.58",
"value": "Nebraska Cornhuskers_sentence_2"
},
{
"operation": "finish",
"time": "84.674",
"value": "finish"
}
]
}
|
11563
|
The Leesburg International Airport, formerly known as Leesburg Regional Airport, has not been designated as a "User Fee" airport by the U.S. Customs and Border Protection, Department of Homeland Security, and as such, it can serve as the first port of entry for aircraft returning to the continental United States from a foreign country.
|
{
"evidence": [
"Leesburg International Airport_sentence_5"
],
"label": "REFUTE"
}
|
feverous
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nThe Leesburg International Airport, formerly known as Leesburg Regional Airport, has not been designated as a \"User Fee\" airport by the U.S. Customs and Border Protection, Department of Homeland Security, and as such, it can serve as the first port of entry for aircraft returning to the continental United States from a foreign country.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Leesburg International Airport_sentence_5"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 7811,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.501",
"value": "?search="
},
{
"operation": "search",
"time": "26.066",
"value": "Leesburg International Airport"
},
{
"operation": "Now on",
"time": "26.961",
"value": "Leesburg International Airport"
},
{
"operation": "Page search",
"time": "38.787",
"value": "regional"
},
{
"operation": "Highlighting",
"time": "42.104",
"value": "Leesburg International Airport_sentence_2"
},
{
"operation": "Page search",
"time": "47.539",
"value": "user fee"
},
{
"operation": "Highlighting",
"time": "52.343",
"value": "Leesburg International Airport_sentence_5"
},
{
"operation": "Highlighting deleted",
"time": "58.033",
"value": "Leesburg International Airport_sentence_2"
},
{
"operation": "finish",
"time": "81.043",
"value": "finish"
}
]
}
|
11858
|
Leik Myrabo, who first had the idea for laser-propelled Lightcraft, a space- or air-vehicle driven by beam-powered propulsion, in the early 1980s, while working on "Star Wars" anti-missile technology, co-wrote in 1985, with SF author Dean Ing, a popularization of this and other unconventional propulsion concepts in "The Future of Flight", published by Baen.
|
{
"evidence": [
"Leik Myrabo_sentence_3",
"Lightcraft_sentence_0",
"Leik Myrabo_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:\nLeik Myrabo, who first had the idea for laser-propelled Lightcraft, a space- or air-vehicle driven by beam-powered propulsion, in the early 1980s, while working on \"Star Wars\" anti-missile technology, co-wrote in 1985, with SF author Dean Ing, a popularization of this and other unconventional propulsion concepts in \"The Future of Flight\", published by Baen.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Leik Myrabo_sentence_3",
"Lightcraft_sentence_0",
"Leik Myrabo_sentence_4"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 7812,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.131",
"value": "?search="
},
{
"operation": "search",
"time": "93.699",
"value": "Leik Myrabo"
},
{
"operation": "Now on",
"time": "94.695",
"value": "Leik Myrabo"
},
{
"operation": "Highlighting",
"time": "114.842",
"value": "Leik Myrabo_sentence_3"
},
{
"operation": "hyperlink",
"time": "131.987",
"value": "lightcraft"
},
{
"operation": "Now on",
"time": "132.913",
"value": "Lightcraft"
},
{
"operation": "Highlighting",
"time": "135.675",
"value": "Lightcraft_sentence_0"
},
{
"operation": "back-button-clicked",
"time": "137.34",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "137.45",
"value": "Leik Myrabo"
},
{
"operation": "Highlighting",
"time": "202.706",
"value": "Leik Myrabo_sentence_4"
},
{
"operation": "finish",
"time": "212.69",
"value": "finish"
}
]
}
|
94702
|
The source of the water for the fountain of Piazza Colonna was only 67 feet above sea level; like the Trevi Fountain, the Fontana della Barcaccia, and the fountains of Piazza Navona, all connected to the Acqua Vergine, the fountain of Piazza Colonna, built by Della Porta, was only set to jet water high into the air.
|
{
"evidence": [
"Fontana di Piazza Colonna_sentence_6"
],
"label": "REFUTE"
}
|
feverous
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nThe source of the water for the fountain of Piazza Colonna was only 67 feet above sea level; like the Trevi Fountain, the Fontana della Barcaccia, and the fountains of Piazza Navona, all connected to the Acqua Vergine, the fountain of Piazza Colonna, built by Della Porta, was only set to jet water high into the air.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Fontana di Piazza Colonna_sentence_6"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 7813,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.441",
"value": "?search="
},
{
"operation": "search",
"time": "31.775",
"value": "Contains...fountain of Piazza Colonna"
},
{
"operation": "Now on",
"time": "33.557",
"value": "undefined"
},
{
"operation": "hyperlink",
"time": "39.655",
"value": "Fontana di Piazza Colonna"
},
{
"operation": "Now on",
"time": "40.391",
"value": "Fontana di Piazza Colonna"
},
{
"operation": "Highlighting",
"time": "97.364",
"value": "Fontana di Piazza Colonna_sentence_6"
},
{
"operation": "finish",
"time": "115.13",
"value": "finish"
}
]
}
|
exfever_3118
|
Perth is part of the South West Land Division of Western Australia, with most of the metropolitan area on the geographic feature which contains the Swan River as it travels west to the Pacific Ocean.
|
{
"evidence": [
"Perth",
"Swan Coastal Plain"
],
"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:\nPerth is part of the South West Land Division of Western Australia, with most of the metropolitan area on the geographic feature which contains the Swan River as it travels west to the Pacific Ocean.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Perth",
"Swan Coastal Plain"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 7814,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
2115
|
Found only in North America, the Habrosyne violacea is a moth that has a wingspan of about 38 mm, which is found by measuring the distance between the center of the thorax to the apex of the wing doubled.
|
{
"evidence": [
"Habrosyne violacea_sentence_1"
],
"label": "REFUTE"
}
|
feverous
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nFound only in North America, the Habrosyne violacea is a moth that has a wingspan of about 38 mm, which is found by measuring the distance between the center of the thorax to the apex of the wing doubled.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Habrosyne violacea_sentence_1"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 7815,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "search",
"time": "8.616",
"value": "Habrosyne violacea"
},
{
"operation": "Highlighting",
"time": "14.42",
"value": "Habrosyne violacea_sentence_1"
},
{
"operation": "Highlighting",
"time": "26.458",
"value": "Habrosyne violacea_sentence_2"
},
{
"operation": "hyperlink",
"time": "27.107",
"value": "forewings"
},
{
"operation": "Page search",
"time": "40.238",
"value": "apex"
},
{
"operation": "finish",
"time": "50.466",
"value": "finish"
}
]
}
|
exfever_4234
|
John D. Rockefeller was an Italian businessman born into a large family in a geographic region consisting of the portion of New York State lying north of the largest metropolitan area in the world by urban landmass.
|
{
"evidence": [
"John D. Rockefeller",
"Upstate New York",
"New York metropolitan area"
],
"label": "REFUTE"
}
|
exfever
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nJohn D. Rockefeller was an Italian businessman born into a large family in a geographic region consisting of the portion of New York State lying north of the largest metropolitan area in the world by urban landmass.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"John D. Rockefeller",
"Upstate New York",
"New York metropolitan area"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 7816,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_6213
|
Seven Days in May is a 1964 film written by an American screenwriter best known for his anthology television series, The Twilight Zone.
|
{
"evidence": [
"Seven Days in May",
"Rod Serling"
],
"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 film written by an American screenwriter best known for his anthology television series, The Twilight Zone.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Seven Days in May",
"Rod Serling"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 7817,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_406
|
Friendships often last longer than romantic relationships.
|
{
"evidence": [
"Friendship",
"Anthropology"
],
"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:\nFriendships often last longer than romantic relationships.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Friendship",
"Anthropology"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 7818,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_8538
|
Chelsea F.C. is the third professional club based in Fulham, West London.
|
{
"evidence": [
"Chelsea F.C.",
"Premier League",
"English football league system"
],
"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:\nChelsea F.C. is the third professional club based in Fulham, West London.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Chelsea F.C.",
"Premier League",
"English football league system"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 7819,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
52483
|
In 1892, the Royal Fusiliers moved to Woolwich, and in 1894, it was posted to Belfast, Ireland.
|
{
"evidence": [
"Royal Fusiliers_sentence_59"
],
"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 1892, the Royal Fusiliers moved to Woolwich, and in 1894, it was posted to Belfast, Ireland.\n",
"role": "user"
}
] |
Numerical Reasoning
|
{
"ground_truth": {
"evidence": [
"Royal Fusiliers_sentence_59"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 7820,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.302",
"value": "?search="
},
{
"operation": "search",
"time": "10.699",
"value": "Royal Fusiliers"
},
{
"operation": "Now on",
"time": "12.83",
"value": "Royal Fusiliers"
},
{
"operation": "Page search",
"time": "23.846",
"value": "Woolwich"
},
{
"operation": "Highlighting",
"time": "42.701",
"value": "Royal Fusiliers_sentence_58"
},
{
"operation": "Highlighting",
"time": "69.961",
"value": "Royal Fusiliers_sentence_59"
},
{
"operation": "Highlighting deleted",
"time": "75.425",
"value": "Royal Fusiliers_sentence_58"
},
{
"operation": "finish",
"time": "100.281",
"value": "finish"
}
]
}
|
4822
|
Renato Ziggiotti failed to join the army during World War I, so he just studied medicine in 1917.
|
{
"evidence": [
"Renato Ziggiotti_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:\nRenato Ziggiotti failed to join the army during World War I, so he just studied medicine in 1917.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Renato Ziggiotti_sentence_7"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 7821,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "search",
"time": "7.124",
"value": "Renato Ziggiotti"
},
{
"operation": "Page search",
"time": "15.547",
"value": "world war"
},
{
"operation": "Highlighting",
"time": "21.97",
"value": "Renato Ziggiotti_sentence_7"
},
{
"operation": "finish",
"time": "35.048",
"value": "finish"
}
]
}
|
exfever_2795
|
The Odd Couple II is Walter Matthau's final film role.
|
{
"evidence": [
"Walter Matthau",
"JFK (film)",
"Lee Harvey Oswald"
],
"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 Odd Couple II is Walter Matthau's final film role.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Walter Matthau",
"JFK (film)",
"Lee Harvey Oswald"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 7822,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_4814
|
The Mughal Empire was an early modern Islamic empire in the southern region of Earth's largest and most populous continent.
|
{
"evidence": [
"Mughal Empire",
"South Asia",
"Asia"
],
"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 Mughal Empire was an early modern Islamic empire in the southern region of Earth's largest and most populous continent.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Mughal Empire",
"South Asia",
"Asia"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 7823,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
16990
|
The Wonder Spot was a tourist attraction located off US Route 12 in Lake Delton, Wisconsin and was purchased by Bill Carney in 1988 who said it had as many as 50,000 visitors each year during the mid-1990's.
|
{
"evidence": [
"The Wonder Spot_sentence_1",
"The Wonder Spot_sentence_5",
"The Wonder Spot_sentence_6"
],
"label": "SUPPORT"
}
|
feverous
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nThe Wonder Spot was a tourist attraction located off US Route 12 in Lake Delton, Wisconsin and was purchased by Bill Carney in 1988 who said it had as many as 50,000 visitors each year during the mid-1990's.\n",
"role": "user"
}
] |
Entity Disambiguation
|
{
"ground_truth": {
"evidence": [
"The Wonder Spot_sentence_1",
"The Wonder Spot_sentence_5",
"The Wonder Spot_sentence_6"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 7824,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.719",
"value": "?search="
},
{
"operation": "search",
"time": "16.353",
"value": "The Wonder Spot"
},
{
"operation": "Now on",
"time": "17.188",
"value": "The Wonder Spot"
},
{
"operation": "Highlighting",
"time": "21.896",
"value": "The Wonder Spot_sentence_0"
},
{
"operation": "Highlighting",
"time": "29.767",
"value": "The Wonder Spot_sentence_1"
},
{
"operation": "Highlighting",
"time": "56.424",
"value": "The Wonder Spot_sentence_5"
},
{
"operation": "Highlighting",
"time": "57.43",
"value": "The Wonder Spot_sentence_6"
},
{
"operation": "Highlighting deleted",
"time": "97.63",
"value": "The Wonder Spot_sentence_0"
},
{
"operation": "finish",
"time": "116.794",
"value": "finish"
}
]
}
|
28192
|
Scotland garnered the most number of gold medals at the Commonwealth Rowing Championships (held neither in August nor in May of 2014) with nine gold medals, followed by England with five gold medals.
|
{
"evidence": [
"Commonwealth Rowing Championships_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:\nScotland garnered the most number of gold medals at the Commonwealth Rowing Championships (held neither in August nor in May of 2014) with nine gold medals, followed by England with five gold medals.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Commonwealth Rowing Championships_sentence_7"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 7825,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.676",
"value": "?search="
},
{
"operation": "search",
"time": "17.926",
"value": "Commonwealth Rowing Championships"
},
{
"operation": "Now on",
"time": "18.944",
"value": "Commonwealth Rowing Championships"
},
{
"operation": "Highlighting",
"time": "32.89",
"value": "Commonwealth Rowing Championships_sentence_7"
},
{
"operation": "Highlighting",
"time": "47.369",
"value": "Commonwealth Rowing Championships_sentence_4"
},
{
"operation": "Highlighting deleted",
"time": "49.126",
"value": "Commonwealth Rowing Championships_sentence_4"
},
{
"operation": "finish",
"time": "55.676",
"value": "finish"
}
]
}
|
70654
|
Thanks to premier institutions such as the Indian Institute of Technology, National Institute of Technology Delhi, and University of Delhi Indian Institute of Technology (one of the first institutes to be granted the status of Institute of Eminence), as well as several Industrial training institute (ITI) and industrial training centres, there are different options when pursuing an education in Delhi.
|
{
"evidence": [
"Education in Delhi_sentence_4",
"IIT Delhi_sentence_27",
"Education in Delhi_sentence_14"
],
"label": "SUPPORT"
}
|
feverous
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nThanks to premier institutions such as the Indian Institute of Technology, National Institute of Technology Delhi, and University of Delhi Indian Institute of Technology (one of the first institutes to be granted the status of Institute of Eminence), as well as several Industrial training institute (ITI) and industrial training centres, there are different options when pursuing an education in Delhi.\n",
"role": "user"
}
] |
Multi-hop Reasoning
|
{
"ground_truth": {
"evidence": [
"Education in Delhi_sentence_4",
"IIT Delhi_sentence_27",
"Education in Delhi_sentence_14"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 7826,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.704",
"value": "?search="
},
{
"operation": "search",
"time": "17.18",
"value": "education in delhi"
},
{
"operation": "Now on",
"time": "18.415",
"value": "Education in Delhi"
},
{
"operation": "Highlighting",
"time": "23.015",
"value": "Education in Delhi_sentence_4"
},
{
"operation": "hyperlink",
"time": "38.218",
"value": "Indian Institute of Technology"
},
{
"operation": "Now on",
"time": "41.651",
"value": "IIT Delhi"
},
{
"operation": "Highlighting",
"time": "67.698",
"value": "IIT Delhi_sentence_27"
},
{
"operation": "back-button-clicked",
"time": "72.697",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "72.801",
"value": "Education in Delhi"
},
{
"operation": "Highlighting",
"time": "103.113",
"value": "Education in Delhi_sentence_14"
},
{
"operation": "finish",
"time": "108.955",
"value": "finish"
}
]
}
|
47191
|
Internet Key Exchange (is the protocol to set up a security ) negotiated key material is then given to the IPsec stack, this could be an AES key.
|
{
"evidence": [
"Internet Key Exchange_sentence_0",
"Internet Key Exchange_sentence_12",
"Internet Key Exchange_sentence_13"
],
"label": "SUPPORT"
}
|
feverous
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nInternet Key Exchange (is the protocol to set up a security ) negotiated key material is then given to the IPsec stack, this could be an AES key.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Internet Key Exchange_sentence_0",
"Internet Key Exchange_sentence_12",
"Internet Key Exchange_sentence_13"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 7827,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.275",
"value": "?search="
},
{
"operation": "search",
"time": "11.055",
"value": "Internet Key Exchange"
},
{
"operation": "Now on",
"time": "12.348",
"value": "Internet Key Exchange"
},
{
"operation": "Highlighting",
"time": "16.558",
"value": "Internet Key Exchange_sentence_0"
},
{
"operation": "Page search",
"time": "31.612",
"value": "IPsec"
},
{
"operation": "Highlighting",
"time": "42.622",
"value": "Internet Key Exchange_sentence_12"
},
{
"operation": "Highlighting",
"time": "43.51",
"value": "Internet Key Exchange_sentence_13"
},
{
"operation": "finish",
"time": "49.255",
"value": "finish"
}
]
}
|
43446
|
Mike Budenholzer went on to lead the Hawks to a franchise record 60 wins, he was named the recipient of the Red Auerbach Trophy named after the Bolton Celtics head coach but never held a higher position.
|
{
"evidence": [
"Mike Budenholzer_sentence_17"
],
"label": "REFUTE"
}
|
feverous
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nMike Budenholzer went on to lead the Hawks to a franchise record 60 wins, he was named the recipient of the Red Auerbach Trophy named after the Bolton Celtics head coach but never held a higher position.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Mike Budenholzer_sentence_17"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 7828,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.781",
"value": "?search="
},
{
"operation": "search",
"time": "7.792",
"value": "Mike Budenholzer"
},
{
"operation": "Now on",
"time": "9.016",
"value": "Mike Budenholzer"
},
{
"operation": "Page search",
"time": "23.123",
"value": "60"
},
{
"operation": "Highlighting",
"time": "26.308",
"value": "Mike Budenholzer_sentence_15"
},
{
"operation": "Highlighting",
"time": "30.808",
"value": "Mike Budenholzer_sentence_16"
},
{
"operation": "hyperlink",
"time": "32.723",
"value": "Red Auerbach"
},
{
"operation": "Now on",
"time": "38.427",
"value": "Red Auerbach"
},
{
"operation": "back-button-clicked",
"time": "41.019",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "41.166",
"value": "Mike Budenholzer"
},
{
"operation": "hyperlink",
"time": "56.132",
"value": "Red Auerbach"
},
{
"operation": "Now on",
"time": "56.658",
"value": "Red Auerbach"
},
{
"operation": "Page search",
"time": "68.529",
"value": "trophy"
},
{
"operation": "Highlighting",
"time": "78.213",
"value": "Red Auerbach_sentence_11"
},
{
"operation": "Highlighting",
"time": "78.215",
"value": "Red Auerbach_title"
},
{
"operation": "Highlighting",
"time": "86.386",
"value": "Red Auerbach_sentence_1"
},
{
"operation": "back-button-clicked",
"time": "97.194",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "97.346",
"value": "Mike Budenholzer"
},
{
"operation": "Highlighting",
"time": "183.282",
"value": "Mike Budenholzer_sentence_17"
},
{
"operation": "Highlighting deleted",
"time": "184.357",
"value": "Mike Budenholzer_sentence_16"
},
{
"operation": "Highlighting deleted",
"time": "184.836",
"value": "Mike Budenholzer_sentence_15"
},
{
"operation": "Highlighting",
"time": "190.629",
"value": "Mike Budenholzer_sentence_17"
},
{
"operation": "finish",
"time": "194.864",
"value": "finish"
}
]
}
|
exfever_4669
|
Legion is an American superhero television series based on a character in a brand name and primary imprint of Marvel Worldwide Inc.
|
{
"evidence": [
"Legion (TV series)",
"Marvel Comics"
],
"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:\nLegion is an American superhero television series based on a character in a brand name and primary imprint of Marvel Worldwide Inc.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Legion (TV series)",
"Marvel Comics"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 7829,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
86175
|
In the novel Five Smooth Stones, David chooses to stay in New Orleans after falling out with Sara Kent.
|
{
"evidence": [
"Five Smooth Stones_sentence_6",
"Five Smooth Stones_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:\nIn the novel Five Smooth Stones, David chooses to stay in New Orleans after falling out with Sara Kent.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Five Smooth Stones_sentence_6",
"Five Smooth Stones_sentence_4"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 7830,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.539",
"value": "undefined"
},
{
"operation": "search",
"time": "15.859",
"value": "Five Smooth Stones"
},
{
"operation": "Now on",
"time": "16.631",
"value": "Five Smooth Stones"
},
{
"operation": "Page search",
"time": "24.635",
"value": "new orleans"
},
{
"operation": "Highlighting",
"time": "27.346",
"value": "Five Smooth Stones_sentence_6"
},
{
"operation": "hyperlink",
"time": "34.416",
"value": "New Orleans"
},
{
"operation": "Now on",
"time": "35.741",
"value": "New Orleans"
},
{
"operation": "back-button-clicked",
"time": "36.779",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "36.897",
"value": "Five Smooth Stones"
},
{
"operation": "Page search",
"time": "46.064",
"value": "new orleans"
},
{
"operation": "Page search",
"time": "64.639",
"value": "sara kent"
},
{
"operation": "Highlighting",
"time": "67.136",
"value": "Five Smooth Stones_sentence_4"
},
{
"operation": "Highlighting deleted",
"time": "76.256",
"value": "Five Smooth Stones_sentence_4"
},
{
"operation": "Page search",
"time": "86.754",
"value": "sara"
},
{
"operation": "search",
"time": "185.331",
"value": "Contains...In the novel Five Smooth Stones, David chooses to stay in New Orleans after falling out with Sara Kent."
},
{
"operation": "Now on",
"time": "187.347",
"value": "undefined"
},
{
"operation": "search",
"time": "204.877",
"value": "Contains...David chooses to stay in New Orleans after falling out with Sara Kent"
},
{
"operation": "Now on",
"time": "205.755",
"value": "undefined"
},
{
"operation": "back-button-clicked",
"time": "243.889",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "244.36",
"value": "undefined"
},
{
"operation": "back-button-clicked",
"time": "245.391",
"value": "back-button-clicked"
},
{
"operation": "back-button-clicked",
"time": "245.395",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "245.448",
"value": "Five Smooth Stones"
},
{
"operation": "Highlighting",
"time": "263.922",
"value": "Five Smooth Stones_sentence_4"
},
{
"operation": "finish",
"time": "281.152",
"value": "finish"
}
]
}
|
exfever_7357
|
The film The Fifth Element's central plot involves the survival of planet Earth, which becomes the responsibility of Korben Dallas, a taxicab driver and former military unit trained to conduct special operations, military activities conducted by specially designated, organized, trained, and equipped forces that were defined as by The North Atlantic Treaty Organization, an intergovernmental military alliance between 30 member states – 28 European states, the United States, and Canada.
|
{
"evidence": [
"The Fifth Element",
"Special forces",
"NATO"
],
"label": "SUPPORT"
}
|
exfever
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nThe film The Fifth Element's central plot involves the survival of planet Earth, which becomes the responsibility of Korben Dallas, a taxicab driver and former military unit trained to conduct special operations, military activities conducted by specially designated, organized, trained, and equipped forces that were defined as by The North Atlantic Treaty Organization, an intergovernmental military alliance between 30 member states – 28 European states, the United States, and Canada.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"The Fifth Element",
"Special forces",
"NATO"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 7831,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
51410
|
Sun Shoubo who is a Chinese football player who currently plays for Wuhan Zall which is a professional Chinese football club that currently participates in the Chinese Super League.
|
{
"evidence": [
"Sun Shoubo_sentence_1",
"Wuhan Zall F.C._sentence_0"
],
"label": "SUPPORT"
}
|
feverous
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nSun Shoubo who is a Chinese football player who currently plays for Wuhan Zall which is a professional Chinese football club that currently participates in the Chinese Super League.\n",
"role": "user"
}
] |
Multi-hop Reasoning
|
{
"ground_truth": {
"evidence": [
"Sun Shoubo_sentence_1",
"Wuhan Zall F.C._sentence_0"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 7832,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.583",
"value": "?search="
},
{
"operation": "search",
"time": "10.257",
"value": "Sun Shoubo"
},
{
"operation": "Now on",
"time": "11.366",
"value": "Sun Shoubo"
},
{
"operation": "Highlighting",
"time": "22.376",
"value": "Sun Shoubo_sentence_1"
},
{
"operation": "hyperlink",
"time": "24.702",
"value": "Wuhan Zall"
},
{
"operation": "Now on",
"time": "26.263",
"value": "Wuhan Zall"
},
{
"operation": "Highlighting",
"time": "27.931",
"value": "Wuhan Zall F.C._sentence_0"
},
{
"operation": "finish",
"time": "34.854",
"value": "finish"
}
]
}
|
77417
|
Several players have won the Bob Skilton medal more than once, however, Bob Skilton himself was never a recipient.
|
{
"evidence": [
"Bob Skilton Medal_sentence_1"
],
"label": "REFUTE"
}
|
feverous
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nSeveral players have won the Bob Skilton medal more than once, however, Bob Skilton himself was never a recipient.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Bob Skilton Medal_sentence_1"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 7833,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.863",
"value": "undefined"
},
{
"operation": "search",
"time": "12.213",
"value": "Bob Skilton Medal"
},
{
"operation": "Now on",
"time": "13.709",
"value": "Bob Skilton Medal"
},
{
"operation": "Highlighting",
"time": "21.918",
"value": "Bob Skilton Medal_sentence_1"
},
{
"operation": "finish",
"time": "29.204",
"value": "finish"
}
]
}
|
exfever_1094
|
Ken Russell, a British film director, has directed many narrative films with a running time long enough to be considered the principal or sole presentation in a commercial entertainment program.
|
{
"evidence": [
"Ken Russell",
"Feature 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:\nKen Russell, a British film director, has directed many narrative films with a running time long enough to be considered the principal or sole presentation in a commercial entertainment program.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Ken Russell",
"Feature film"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 7834,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
89357
|
The New Zealand national rugby team, whose colors are black and white and known as the Kiwis, named Anthony Cherrington but did not play him, and the next year he was charged for assaulting his ex and got community service.
|
{
"evidence": [
"New Zealand national rugby league team_sentence_4",
"New Zealand national rugby league team_sentence_3",
"Anthony Cherrington_sentence_5",
"Anthony Cherrington_sentence_6",
"Anthony Cherrington_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:\nThe New Zealand national rugby team, whose colors are black and white and known as the Kiwis, named Anthony Cherrington but did not play him, and the next year he was charged for assaulting his ex and got community service.\n",
"role": "user"
}
] |
Multi-hop Reasoning
|
{
"ground_truth": {
"evidence": [
"New Zealand national rugby league team_sentence_4",
"New Zealand national rugby league team_sentence_3",
"Anthony Cherrington_sentence_5",
"Anthony Cherrington_sentence_6",
"Anthony Cherrington_sentence_8"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 7835,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.823",
"value": "?search="
},
{
"operation": "search",
"time": "12.075",
"value": "New Zealand national rugby team"
},
{
"operation": "Now on",
"time": "12.741",
"value": "New Zealand national rugby team"
},
{
"operation": "search",
"time": "36.435",
"value": "the kiwis"
},
{
"operation": "Now on",
"time": "37.346",
"value": "undefined"
},
{
"operation": "back-button-clicked",
"time": "49.053",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "49.123",
"value": "New Zealand national rugby team"
},
{
"operation": "hyperlink",
"time": "52.333",
"value": "New Zealand national rugby league team"
},
{
"operation": "Now on",
"time": "54.895",
"value": "New Zealand national rugby league team"
},
{
"operation": "Highlighting",
"time": "58.954",
"value": "New Zealand national rugby league team_sentence_4"
},
{
"operation": "Highlighting",
"time": "62.475",
"value": "New Zealand national rugby league team_sentence_3"
},
{
"operation": "search",
"time": "81.878",
"value": "Anthony Cherrington"
},
{
"operation": "Now on",
"time": "82.433",
"value": "Anthony Cherrington"
},
{
"operation": "Highlighting",
"time": "103.017",
"value": "Anthony Cherrington_sentence_5"
},
{
"operation": "Highlighting",
"time": "162.434",
"value": "Anthony Cherrington_sentence_6"
},
{
"operation": "Highlighting deleted",
"time": "163.288",
"value": "Anthony Cherrington_sentence_6"
},
{
"operation": "Highlighting",
"time": "190.048",
"value": "Anthony Cherrington_sentence_7"
},
{
"operation": "Highlighting deleted",
"time": "190.543",
"value": "Anthony Cherrington_sentence_7"
},
{
"operation": "Highlighting",
"time": "191.713",
"value": "Anthony Cherrington_sentence_6"
},
{
"operation": "Highlighting",
"time": "192.239",
"value": "Anthony Cherrington_sentence_7"
},
{
"operation": "Highlighting deleted",
"time": "194.739",
"value": "Anthony Cherrington_sentence_7"
},
{
"operation": "Highlighting",
"time": "195.354",
"value": "Anthony Cherrington_sentence_8"
},
{
"operation": "hyperlink",
"time": "220.315",
"value": "New Zealand training squad"
},
{
"operation": "Now on",
"time": "220.825",
"value": "New Zealand national rugby league team#2008 World Cup Squad"
},
{
"operation": "back-button-clicked",
"time": "294.715",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "294.787",
"value": "Anthony Cherrington"
},
{
"operation": "hyperlink",
"time": "307.111",
"value": "New Zealand training squad"
},
{
"operation": "Now on",
"time": "307.513",
"value": "New Zealand national rugby league team#2008 World Cup Squad"
},
{
"operation": "back-button-clicked",
"time": "314.689",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "314.799",
"value": "Anthony Cherrington"
},
{
"operation": "hyperlink",
"time": "395.871",
"value": "2008 Rugby League World Cup"
},
{
"operation": "Now on",
"time": "397.04",
"value": "2008 Rugby League World Cup"
},
{
"operation": "back-button-clicked",
"time": "420.353",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "420.464",
"value": "Anthony Cherrington"
},
{
"operation": "finish",
"time": "436.615",
"value": "finish"
}
]
}
|
exfever_7497
|
Shawn Michaels performed in non-wrestling roles for two years before returning to the ring for a match in his own professional wrestling training school that has produced many well-known wrestlers, including Manny Pacquiao.
|
{
"evidence": [
"Shawn Michaels",
"Texas Wrestling Academy"
],
"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:\nShawn Michaels performed in non-wrestling roles for two years before returning to the ring for a match in his own professional wrestling training school that has produced many well-known wrestlers, including Manny Pacquiao.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Shawn Michaels",
"Texas Wrestling Academy"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 7836,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_6802
|
Roger Penrose is a Junior Professor of Mathematics at a collegiate research university in Oxford, England.
|
{
"evidence": [
"Roger Penrose",
"University of Oxford"
],
"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:\nRoger Penrose is a Junior Professor of Mathematics at a collegiate research university in Oxford, England.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Roger Penrose",
"University of Oxford"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 7837,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
54943
|
Western Australia that has been held by multiple people such as William Angwin (who was Deputy Premier of Western Australia at the time) in 1924, and Frank Troy in 1933.
|
{
"evidence": [
"William Angwin_sentence_2",
"Frank Troy_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:\nWestern Australia that has been held by multiple people such as William Angwin (who was Deputy Premier of Western Australia at the time) in 1924, and Frank Troy in 1933.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"William Angwin_sentence_2",
"Frank Troy_sentence_3"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 7838,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.168",
"value": "?search="
},
{
"operation": "search",
"time": "45.799",
"value": "western australia deputy premier"
},
{
"operation": "Now on",
"time": "48.222",
"value": "undefined"
},
{
"operation": "hyperlink",
"time": "50.734",
"value": "Deputy Premier of Western Australia"
},
{
"operation": "Now on",
"time": "51.811",
"value": "Deputy Premier of Western Australia"
},
{
"operation": "search",
"time": "109.872",
"value": "western australia"
},
{
"operation": "Now on",
"time": "118.431",
"value": "Western Australia"
},
{
"operation": "search",
"time": "143.223",
"value": "William Angwin"
},
{
"operation": "Now on",
"time": "144.047",
"value": "William Angwin"
},
{
"operation": "Highlighting",
"time": "153.021",
"value": "William Angwin_sentence_2"
},
{
"operation": "search",
"time": "216.432",
"value": "Frank Troy"
},
{
"operation": "Now on",
"time": "217.412",
"value": "Frank Troy"
},
{
"operation": "Highlighting",
"time": "258.681",
"value": "Frank Troy_header_cell_0_10_0"
},
{
"operation": "Highlighting deleted",
"time": "259.692",
"value": "Frank Troy_header_cell_0_10_0"
},
{
"operation": "Highlighting",
"time": "270.503",
"value": "Frank Troy_sentence_3"
},
{
"operation": "finish",
"time": "278.547",
"value": "finish"
}
]
}
|
29975
|
The article reported that Charles Dougherty (former Duquesne University President) praised students who live off-campus as the ideal lifestyle for students and described their situation as "a preparation for being an adult." The comments gained popularity in the campus community and received coverage from national outlets such as The New York Times.
|
{
"evidence": [
"Charles J. Dougherty_sentence_11",
"Charles J. Dougherty_sentence_12",
"Charles J. Dougherty_sentence_15"
],
"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 article reported that Charles Dougherty (former Duquesne University President) praised students who live off-campus as the ideal lifestyle for students and described their situation as \"a preparation for being an adult.\" The comments gained popularity in the campus community and received coverage from national outlets such as The New York Times.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Charles J. Dougherty_sentence_11",
"Charles J. Dougherty_sentence_12",
"Charles J. Dougherty_sentence_15"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 7839,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.248",
"value": "?search="
},
{
"operation": "search",
"time": "45.584",
"value": "Charles Dougherty"
},
{
"operation": "Now on",
"time": "46.358",
"value": "Charles Dougherty"
},
{
"operation": "hyperlink",
"time": "56.417",
"value": "Charles J. Dougherty"
},
{
"operation": "Now on",
"time": "57.048",
"value": "Charles J. Dougherty"
},
{
"operation": "hyperlink",
"time": "108.11",
"value": "Duquesne University"
},
{
"operation": "Now on",
"time": "109.505",
"value": "Duquesne University"
},
{
"operation": "Page search",
"time": "123.171",
"value": "Charles Dougherty"
},
{
"operation": "search",
"time": "173.607",
"value": "Contains... The comments gained popularity in the campus community and received coverage from national outlets such as The New York Times."
},
{
"operation": "Now on",
"time": "177.001",
"value": "undefined"
},
{
"operation": "search",
"time": "200.75",
"value": "Contains...praised students who live off-campus as the ideal lifestyle for students and described their situation as \"a preparation for being an adult"
},
{
"operation": "Now on",
"time": "202.146",
"value": "undefined"
},
{
"operation": "search",
"time": "242.71",
"value": "Charles Dougherty praised students who live off-campus as the ideal lifestyle for students and described their situation as a preparation for being an adult."
},
{
"operation": "Now on",
"time": "243.453",
"value": "undefined"
},
{
"operation": "search",
"time": "270.542",
"value": "Charles Dougherty"
},
{
"operation": "Now on",
"time": "271.699",
"value": "Charles Dougherty"
},
{
"operation": "hyperlink",
"time": "276.995",
"value": "Charles J. Dougherty"
},
{
"operation": "Now on",
"time": "277.785",
"value": "Charles J. Dougherty"
},
{
"operation": "Highlighting",
"time": "320.478",
"value": "Charles J. Dougherty_sentence_11"
},
{
"operation": "Highlighting",
"time": "321.692",
"value": "Charles J. Dougherty_sentence_12"
},
{
"operation": "Highlighting",
"time": "323.678",
"value": "Charles J. Dougherty_sentence_14"
},
{
"operation": "Highlighting",
"time": "327.381",
"value": "Charles J. Dougherty_sentence_15"
},
{
"operation": "finish",
"time": "346.418",
"value": "finish"
}
]
}
|
45924
|
Eduard Wölfflin, who studied at the Universities of Basel and Göttingen from 1848 to 1854, worked as a librarian at the University of Basel following graduation before becoming a schoolteacher in Winterthur (the country's sixth largest city by population) and eventually a full professor at the University of Zurich..
|
{
"evidence": [
"Eduard Wölfflin_sentence_2",
"Eduard Wölfflin_sentence_3",
"Eduard Wölfflin_sentence_4",
"Winterthur_sentence_2",
"Eduard Wölfflin_sentence_5"
],
"label": "SUPPORT"
}
|
feverous
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nEduard Wölfflin, who studied at the Universities of Basel and Göttingen from 1848 to 1854, worked as a librarian at the University of Basel following graduation before becoming a schoolteacher in Winterthur (the country's sixth largest city by population) and eventually a full professor at the University of Zurich..\n",
"role": "user"
}
] |
Multi-hop Reasoning
|
{
"ground_truth": {
"evidence": [
"Eduard Wölfflin_sentence_2",
"Eduard Wölfflin_sentence_3",
"Eduard Wölfflin_sentence_4",
"Winterthur_sentence_2",
"Eduard Wölfflin_sentence_5"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 7840,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.771",
"value": "?search="
},
{
"operation": "search",
"time": "9.779",
"value": "Eduard Wölfflin"
},
{
"operation": "Now on",
"time": "10.501",
"value": "Eduard Wölfflin"
},
{
"operation": "Highlighting",
"time": "25.886",
"value": "Eduard Wölfflin_sentence_3"
},
{
"operation": "Highlighting deleted",
"time": "52.197",
"value": "Eduard Wölfflin_sentence_3"
},
{
"operation": "Highlighting",
"time": "54.174",
"value": "Eduard Wölfflin_sentence_2"
},
{
"operation": "Highlighting",
"time": "55.795",
"value": "Eduard Wölfflin_sentence_3"
},
{
"operation": "Highlighting",
"time": "63.26",
"value": "Eduard Wölfflin_sentence_4"
},
{
"operation": "hyperlink",
"time": "73.002",
"value": "Winterthur"
},
{
"operation": "Now on",
"time": "74.508",
"value": "Winterthur"
},
{
"operation": "Highlighting",
"time": "86.35",
"value": "Winterthur_sentence_2"
},
{
"operation": "back-button-clicked",
"time": "100.055",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "100.175",
"value": "Eduard Wölfflin"
},
{
"operation": "Highlighting",
"time": "109.446",
"value": "Eduard Wölfflin_sentence_5"
},
{
"operation": "finish",
"time": "135.206",
"value": "finish"
}
]
}
|
exfever_5340
|
Steve Winwood is widely regarded as one of the most influential musician of all time.
|
{
"evidence": [
"Blind Faith",
"Eric Clapton"
],
"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:\nSteve Winwood is widely regarded as one of the most influential musician of all time.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Blind Faith",
"Eric Clapton"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 7841,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
88992
|
Jimmy McIlvenny scored 20 goals in 25 games in 1915–16 season and scored another two hat-tricks the following season among 26 goals from 32 games, before scoring another 12 from 28 in 1917–18.
|
{
"evidence": [
"Jimmy McIlvenny_sentence_6",
"Jimmy McIlvenny_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:\nJimmy McIlvenny scored 20 goals in 25 games in 1915–16 season and scored another two hat-tricks the following season among 26 goals from 32 games, before scoring another 12 from 28 in 1917–18.\n",
"role": "user"
}
] |
Numerical Reasoning
|
{
"ground_truth": {
"evidence": [
"Jimmy McIlvenny_sentence_6",
"Jimmy McIlvenny_sentence_7"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 7842,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.891",
"value": "?search="
},
{
"operation": "search",
"time": "50.757",
"value": "Jimmy McIlvenny"
},
{
"operation": "Now on",
"time": "51.774",
"value": "Jimmy McIlvenny"
},
{
"operation": "Highlighting",
"time": "239.649",
"value": "Jimmy McIlvenny_sentence_6"
},
{
"operation": "Highlighting",
"time": "244.279",
"value": "Jimmy McIlvenny_sentence_7"
},
{
"operation": "finish",
"time": "260.286",
"value": "finish"
}
]
}
|
19978
|
Hafiz Hayat was advised to settle across the Chenab River which is a major river that flows around India and Pakistan and rises in the upper Himalayas and flows through the Jammu region before flowing into the Indus River.
|
{
"evidence": [
"Hafiz Hayat_sentence_3",
"Chenab River_sentence_5",
"Chenab River_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:\nHafiz Hayat was advised to settle across the Chenab River which is a major river that flows around India and Pakistan and rises in the upper Himalayas and flows through the Jammu region before flowing into the Indus River.\n",
"role": "user"
}
] |
Multi-hop Reasoning
|
{
"ground_truth": {
"evidence": [
"Hafiz Hayat_sentence_3",
"Chenab River_sentence_5",
"Chenab River_sentence_6"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 7843,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.939",
"value": "?search="
},
{
"operation": "search",
"time": "10.574",
"value": "Hafiz Hayat"
},
{
"operation": "Now on",
"time": "11.416",
"value": "Hafiz Hayat"
},
{
"operation": "Page search",
"time": "24.502",
"value": "Chenab River"
},
{
"operation": "Page search",
"time": "24.706",
"value": "Chenab River"
},
{
"operation": "Page search",
"time": "24.785",
"value": "Chenab River"
},
{
"operation": "Highlighting",
"time": "27.144",
"value": "Hafiz Hayat_sentence_3"
},
{
"operation": "hyperlink",
"time": "39.286",
"value": "Chenab River"
},
{
"operation": "Now on",
"time": "40.046",
"value": "Chenab River"
},
{
"operation": "Highlighting",
"time": "52.826",
"value": "Chenab River_sentence_5"
},
{
"operation": "Highlighting",
"time": "69.221",
"value": "Chenab River_sentence_6"
},
{
"operation": "finish",
"time": "177.603",
"value": "finish"
}
]
}
|
exfever_144
|
Stephanie Daley, retitled What She Knew for US television, is a film starring an American writer, director, and actress.
|
{
"evidence": [
"Stephanie Daley",
"Amber Tamblyn"
],
"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:\nStephanie Daley, retitled What She Knew for US television, is a film starring an American writer, director, and actress.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Stephanie Daley",
"Amber Tamblyn"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 7844,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
36300
|
In race two of nine of the 1962 World Championship of Drivers, the Monaco Grand Prix, Bruce McLaren qualified at number one with BRM while Jim Clark qualified at number two with Ferrari.
|
{
"evidence": [
"Bruce McLaren_sentence_22"
],
"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 race two of nine of the 1962 World Championship of Drivers, the Monaco Grand Prix, Bruce McLaren qualified at number one with BRM while Jim Clark qualified at number two with Ferrari.\n",
"role": "user"
}
] |
Multi-hop Reasoning
|
{
"ground_truth": {
"evidence": [
"Bruce McLaren_sentence_22"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 7845,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.762",
"value": "?search="
},
{
"operation": "search",
"time": "27.016",
"value": "1962 World Championship of Drivers"
},
{
"operation": "Now on",
"time": "28.935",
"value": "1962 Formula One season"
},
{
"operation": "Highlighting",
"time": "49.249",
"value": "1962 Formula One season_sentence_1"
},
{
"operation": "Highlighting",
"time": "49.253",
"value": "1962 Formula One season_title"
},
{
"operation": "Highlighting",
"time": "103.648",
"value": "1962 Formula One season_cell_0_2_6"
},
{
"operation": "Highlighting",
"time": "103.653",
"value": "1962 Formula One season_section_1"
},
{
"operation": "Highlighting",
"time": "103.667",
"value": "1962 Formula One season_header_cell_0_2_0"
},
{
"operation": "Highlighting",
"time": "103.668",
"value": "1962 Formula One season_header_cell_0_0_6"
},
{
"operation": "search",
"time": "278.705",
"value": "Bruce McLaren"
},
{
"operation": "Now on",
"time": "280.399",
"value": "Bruce McLaren"
},
{
"operation": "Highlighting",
"time": "339.438",
"value": "Bruce McLaren_sentence_22"
},
{
"operation": "search",
"time": "375.933",
"value": "Jim Clark"
},
{
"operation": "Now on",
"time": "377.817",
"value": "Jim Clark"
},
{
"operation": "finish",
"time": "566.391",
"value": "finish"
}
]
}
|
76033
|
Jacob Zhitomirsky, recruited by Okhrana in 1902, was active in Berlin RSPLP group, reporting its activities to the Russian police, until 1907 when Bolsheviks were expelled from Berlin by German authorities and Zhitomirsky moved to Paris.
|
{
"evidence": [
"Jacob Zhitomirsky_sentence_1",
"Jacob Zhitomirsky_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:\nJacob Zhitomirsky, recruited by Okhrana in 1902, was active in Berlin RSPLP group, reporting its activities to the Russian police, until 1907 when Bolsheviks were expelled from Berlin by German authorities and Zhitomirsky moved to Paris.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Jacob Zhitomirsky_sentence_1",
"Jacob Zhitomirsky_sentence_2"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 7846,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.994",
"value": "undefined"
},
{
"operation": "search",
"time": "33.731",
"value": "Jacob Zhitomirsky,"
},
{
"operation": "Now on",
"time": "36.087",
"value": "undefined"
},
{
"operation": "hyperlink",
"time": "47.859",
"value": "Jacob Zhitomirsky"
},
{
"operation": "Now on",
"time": "49.7",
"value": "Jacob Zhitomirsky"
},
{
"operation": "Highlighting",
"time": "64.465",
"value": "Jacob Zhitomirsky_sentence_0"
},
{
"operation": "Highlighting deleted",
"time": "68.796",
"value": "Jacob Zhitomirsky_sentence_0"
},
{
"operation": "Highlighting",
"time": "69.99",
"value": "Jacob Zhitomirsky_sentence_1"
},
{
"operation": "Highlighting",
"time": "79.571",
"value": "Jacob Zhitomirsky_sentence_2"
},
{
"operation": "finish",
"time": "100.812",
"value": "finish"
}
]
}
|
29819
|
In 1962, Paul Rothchild, a friend of the band who had produced their second album on his own label, Mount Auburn Records, began working for Prestige Records, which reissued the album as Bluegrass And Old Timey Music (1962), and produced a further album on the label, Blue Grass Get Together (with Tex Logan, 1964).
|
{
"evidence": [
"Charles River Valley Boys_sentence_7",
"Charles River Valley Boys_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:\nIn 1962, Paul Rothchild, a friend of the band who had produced their second album on his own label, Mount Auburn Records, began working for Prestige Records, which reissued the album as Bluegrass And Old Timey Music (1962), and produced a further album on the label, Blue Grass Get Together (with Tex Logan, 1964).\n",
"role": "user"
}
] |
Search terms not in claim
|
{
"ground_truth": {
"evidence": [
"Charles River Valley Boys_sentence_7",
"Charles River Valley Boys_sentence_8"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 7847,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.681",
"value": "undefined"
},
{
"operation": "search",
"time": "7.483",
"value": "Paul Rothchild"
},
{
"operation": "Now on",
"time": "8.366",
"value": "Paul A. Rothchild"
},
{
"operation": "search",
"time": "79.104",
"value": "bluegrass and old timey music"
},
{
"operation": "Now on",
"time": "80.408",
"value": "undefined"
},
{
"operation": "hyperlink",
"time": "84.727",
"value": "Charles River Valley Boys"
},
{
"operation": "Now on",
"time": "85.323",
"value": "Charles River Valley Boys"
},
{
"operation": "Highlighting",
"time": "91.882",
"value": "Charles River Valley Boys_sentence_7"
},
{
"operation": "Highlighting",
"time": "101.775",
"value": "Charles River Valley Boys_sentence_8"
},
{
"operation": "finish",
"time": "140.983",
"value": "finish"
}
]
}
|
33184
|
In March 1940, BMI issued their very first contract, which was for "We Could Make Such Beautiful Music Together", with music composed by Katzman and lyrics by Robert Sour (president of Broadcast Music Incorporated (BMI)); which was frequently covered by a wide variety of artists, including vintage recordings by Barbara Lea, Judy Garland, and Liza Minnelli.
|
{
"evidence": [
"Henry Katzman_sentence_5",
"Robert Sour_sentence_0",
"Henry Katzman_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:\nIn March 1940, BMI issued their very first contract, which was for \"We Could Make Such Beautiful Music Together\", with music composed by Katzman and lyrics by Robert Sour (president of Broadcast Music Incorporated (BMI)); which was frequently covered by a wide variety of artists, including vintage recordings by Barbara Lea, Judy Garland, and Liza Minnelli.\n",
"role": "user"
}
] |
Search terms not in claim
|
{
"ground_truth": {
"evidence": [
"Henry Katzman_sentence_5",
"Robert Sour_sentence_0",
"Henry Katzman_sentence_6"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 7848,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.724",
"value": "?search="
},
{
"operation": "search",
"time": "16.18",
"value": "BMI"
},
{
"operation": "Now on",
"time": "17.521",
"value": "BMI"
},
{
"operation": "hyperlink",
"time": "21.414",
"value": "BMI Foundation"
},
{
"operation": "Now on",
"time": "21.906",
"value": "BMI Foundation"
},
{
"operation": "search",
"time": "38.117",
"value": "Contains...We Could Make Such Beautiful Music Together"
},
{
"operation": "Now on",
"time": "41.089",
"value": "undefined"
},
{
"operation": "hyperlink",
"time": "46.188",
"value": "Henry Katzman"
},
{
"operation": "Now on",
"time": "46.798",
"value": "Henry Katzman"
},
{
"operation": "Page search",
"time": "59.476",
"value": "BMI"
},
{
"operation": "Highlighting",
"time": "66.013",
"value": "Henry Katzman_sentence_5"
},
{
"operation": "hyperlink",
"time": "75.814",
"value": "Robert Sour"
},
{
"operation": "Now on",
"time": "76.484",
"value": "Robert Sour"
},
{
"operation": "Highlighting",
"time": "78.838",
"value": "Robert Sour_sentence_0"
},
{
"operation": "back-button-clicked",
"time": "80.621",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "80.71",
"value": "Henry Katzman"
},
{
"operation": "Highlighting",
"time": "86.082",
"value": "Henry Katzman_sentence_6"
},
{
"operation": "finish",
"time": "115.088",
"value": "finish"
}
]
}
|
exfever_4750
|
William Hanna was an American animator who joined an American animation team known for founding DreamWorks.
|
{
"evidence": [
"William Hanna",
"Harman and Ising"
],
"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:\nWilliam Hanna was an American animator who joined an American animation team known for founding DreamWorks.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"William Hanna",
"Harman and Ising"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 7849,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
31605
|
The shooting event of the modern pentathlon at the 1924 Summer Olympics, where Sweden was able to sweep the medals, was participated by shooters from various countries such as Sweden, France, USA, and Denmark and was topped by Otto Olsen from Denmark.
|
{
"evidence": [
"Otto Olsen (sport shooter)_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 shooting event of the modern pentathlon at the 1924 Summer Olympics, where Sweden was able to sweep the medals, was participated by shooters from various countries such as Sweden, France, USA, and Denmark and was topped by Otto Olsen from Denmark.\n",
"role": "user"
}
] |
Search terms not in claim
|
{
"ground_truth": {
"evidence": [
"Otto Olsen (sport shooter)_sentence_0"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 7850,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.689",
"value": "?search="
},
{
"operation": "search",
"time": "33.93",
"value": "The shooting event of the modern pentathlon at the 1924 Summer Olympics"
},
{
"operation": "Now on",
"time": "36.83",
"value": "undefined"
},
{
"operation": "hyperlink",
"time": "43.133",
"value": "1924 Summer Olympics"
},
{
"operation": "Now on",
"time": "43.784",
"value": "1924 Summer Olympics"
},
{
"operation": "search",
"time": "178.887",
"value": "Otto Olsen"
},
{
"operation": "Now on",
"time": "179.47",
"value": "Otto Olsen"
},
{
"operation": "hyperlink",
"time": "182.469",
"value": "Otto Olsen (sport shooter)"
},
{
"operation": "Now on",
"time": "182.904",
"value": "Otto Olsen (sport shooter)"
},
{
"operation": "hyperlink",
"time": "197.465",
"value": "Norwegian"
},
{
"operation": "Now on",
"time": "198.921",
"value": "Norway"
},
{
"operation": "back-button-clicked",
"time": "226.01",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "226.111",
"value": "Otto Olsen (sport shooter)"
},
{
"operation": "Highlighting",
"time": "229.922",
"value": "Otto Olsen (sport shooter)_sentence_0"
},
{
"operation": "search",
"time": "261.12",
"value": "1924 Summer Olympics medal table"
},
{
"operation": "Now on",
"time": "261.823",
"value": "1924 Summer Olympics medal table"
},
{
"operation": "back-button-clicked",
"time": "279.367",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "279.452",
"value": "Otto Olsen (sport shooter)"
},
{
"operation": "finish",
"time": "290.505",
"value": "finish"
}
]
}
|
79851
|
Bram Zeegers testified in court over two days that Holleeder, nicknamed De Neus, and his associates had extorted thousands of dollars from Endstra and also testified about the tense relationship between Holleeder and Endstra before the latter's 2004 murder.
|
{
"evidence": [
"Bram Zeegers_sentence_4",
"Willem Holleeder_sentence_1",
"Bram Zeegers_sentence_5"
],
"label": "SUPPORT"
}
|
feverous
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nBram Zeegers testified in court over two days that Holleeder, nicknamed De Neus, and his associates had extorted thousands of dollars from Endstra and also testified about the tense relationship between Holleeder and Endstra before the latter's 2004 murder.\n",
"role": "user"
}
] |
Multi-hop Reasoning
|
{
"ground_truth": {
"evidence": [
"Bram Zeegers_sentence_4",
"Willem Holleeder_sentence_1",
"Bram Zeegers_sentence_5"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 7851,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.854",
"value": "?search="
},
{
"operation": "search",
"time": "22.079",
"value": "Bram Zeegers"
},
{
"operation": "Now on",
"time": "22.86",
"value": "Bram Zeegers"
},
{
"operation": "Page search",
"time": "52.986",
"value": "Holleeder"
},
{
"operation": "Highlighting",
"time": "57.088",
"value": "Bram Zeegers_sentence_4"
},
{
"operation": "hyperlink",
"time": "61.624",
"value": "Willem Holleeder"
},
{
"operation": "Now on",
"time": "62.326",
"value": "Willem Holleeder"
},
{
"operation": "Highlighting",
"time": "64.408",
"value": "Willem Holleeder_sentence_1"
},
{
"operation": "back-button-clicked",
"time": "66.449",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "66.724",
"value": "Bram Zeegers"
},
{
"operation": "Highlighting",
"time": "74.448",
"value": "Bram Zeegers_sentence_5"
},
{
"operation": "finish",
"time": "81.033",
"value": "finish"
}
]
}
|
77944
|
Glen Clegg was a civil servant and politician from Alberta, Canada, He served in the Legislative Assembly of Alberta from 1986 to 2001.
|
{
"evidence": [
"Glen Clegg_sentence_0",
"Glen Clegg_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:\nGlen Clegg was a civil servant and politician from Alberta, Canada, He served in the Legislative Assembly of Alberta from 1986 to 2001.\n",
"role": "user"
}
] |
Entity Disambiguation
|
{
"ground_truth": {
"evidence": [
"Glen Clegg_sentence_0",
"Glen Clegg_sentence_1"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 7852,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.937",
"value": "?search="
},
{
"operation": "search",
"time": "10.293",
"value": "Glen Clegg"
},
{
"operation": "Now on",
"time": "11.201",
"value": "Glen Clegg"
},
{
"operation": "Highlighting",
"time": "18.514",
"value": "Glen Clegg_sentence_0"
},
{
"operation": "Highlighting",
"time": "21.238",
"value": "Glen Clegg_sentence_1"
},
{
"operation": "Highlighting",
"time": "27.598",
"value": "Glen Clegg_header_cell_0_0_0"
},
{
"operation": "finish",
"time": "64.794",
"value": "finish"
}
]
}
|
exfever_5681
|
Laurence Olivier is the longest actor that played in Romeo and Juliet.
|
{
"evidence": [
"Laurence Olivier",
"Peggy Ashcroft"
],
"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:\nLaurence Olivier is the longest actor that played in Romeo and Juliet.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Laurence Olivier",
"Peggy Ashcroft"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 7853,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_5950
|
Toy Story is a 1995 film written by an American filmmaker best known as the creator of a 1999 TV series, a spinoff from the television series Buffy the Vampire Slayer.
|
{
"evidence": [
"Toy Story",
"Joss Whedon",
"Angel (1999 TV series)"
],
"label": "SUPPORT"
}
|
exfever
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nToy Story is a 1995 film written by an American filmmaker best known as the creator of a 1999 TV series, a spinoff from the television series Buffy the Vampire Slayer.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Toy Story",
"Joss Whedon",
"Angel (1999 TV series)"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 7854,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_8019
|
The Hangover Part III is a 2013 American action comedy film that stars an American actor who has played paper salesman Andy Bernard in the sitcom The Office on an American English-language commercial broadcast television and radio network owned by Comcast.
|
{
"evidence": [
"The Hangover Part III",
"Ed Helms",
"NBC"
],
"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 Hangover Part III is a 2013 American action comedy film that stars an American actor who has played paper salesman Andy Bernard in the sitcom The Office on an American English-language commercial broadcast television and radio network owned by Comcast.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"The Hangover Part III",
"Ed Helms",
"NBC"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 7855,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
40933
|
The story of Bird Box, a 2000 made-for-TV romance film that starred American former actress Yasmine Bleeth, and Gian-Carlo Scandiuzzi, revolves around Robert, an aspiring novelist, and his wife, Claudia, a talented painter, who are about to sign their divorce papers.
|
{
"evidence": [
"Goodbye Casanova_sentence_0",
"Goodbye Casanova_sentence_1",
"Goodbye Casanova_sentence_2",
"Goodbye Casanova_sentence_3"
],
"label": "REFUTE"
}
|
feverous
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nThe story of Bird Box, a 2000 made-for-TV romance film that starred American former actress Yasmine Bleeth, and Gian-Carlo Scandiuzzi, revolves around Robert, an aspiring novelist, and his wife, Claudia, a talented painter, who are about to sign their divorce papers.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Goodbye Casanova_sentence_0",
"Goodbye Casanova_sentence_1",
"Goodbye Casanova_sentence_2",
"Goodbye Casanova_sentence_3"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 7856,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.565",
"value": "undefined"
},
{
"operation": "search",
"time": "10.613",
"value": "Bird Box"
},
{
"operation": "Now on",
"time": "12.03",
"value": "Bird Box"
},
{
"operation": "hyperlink",
"time": "18.655",
"value": "Bird Box (film)"
},
{
"operation": "Now on",
"time": "19.931",
"value": "Bird Box (film)"
},
{
"operation": "search",
"time": "36.555",
"value": "Yasmine Bleeth"
},
{
"operation": "Now on",
"time": "37.715",
"value": "Yasmine Bleeth"
},
{
"operation": "Page search",
"time": "59.248",
"value": "Bird Box"
},
{
"operation": "search",
"time": "84.148",
"value": "Contains...Gian-Carlo Scandiuzzi"
},
{
"operation": "Now on",
"time": "86.079",
"value": "undefined"
},
{
"operation": "hyperlink",
"time": "92.62",
"value": "Goodbye Casanova"
},
{
"operation": "Now on",
"time": "94.112",
"value": "Goodbye Casanova"
},
{
"operation": "Highlighting",
"time": "104.172",
"value": "Goodbye Casanova_sentence_0"
},
{
"operation": "hyperlink",
"time": "108.057",
"value": "Yasmine Bleeth"
},
{
"operation": "Now on",
"time": "108.862",
"value": "Yasmine Bleeth"
},
{
"operation": "Highlighting",
"time": "113.106",
"value": "Yasmine Bleeth_sentence_0"
},
{
"operation": "Highlighting",
"time": "113.108",
"value": "Yasmine Bleeth_title"
},
{
"operation": "back-button-clicked",
"time": "116.339",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "116.611",
"value": "Goodbye Casanova"
},
{
"operation": "Highlighting",
"time": "125.546",
"value": "Goodbye Casanova_sentence_1"
},
{
"operation": "Highlighting",
"time": "128.602",
"value": "Goodbye Casanova_sentence_2"
},
{
"operation": "Highlighting",
"time": "130.594",
"value": "Goodbye Casanova_sentence_3"
},
{
"operation": "finish",
"time": "168.659",
"value": "finish"
}
]
}
|
exfever_5274
|
Lady Gaga is an American singer who signed to a record label with notable artists like a Canadian rapper, often credited as Canada's hip hop ambassador.
|
{
"evidence": [
"Lady Gaga",
"KonLive Distribution",
"Kardinal Offishall"
],
"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:\nLady Gaga is an American singer who signed to a record label with notable artists like a Canadian rapper, often credited as Canada's hip hop ambassador.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Lady Gaga",
"KonLive Distribution",
"Kardinal Offishall"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 7857,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_7339
|
Here Comes the Sun is a song written by an English musician, singer-songwriter, and music and film producer who achieved international fame as the lead guitarist of the Beatles.
|
{
"evidence": [
"Here Comes the Sun",
"George Harrison"
],
"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:\nHere Comes the Sun is a song written by an English musician, singer-songwriter, and music and film producer who achieved international fame as the lead guitarist of the Beatles.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Here Comes the Sun",
"George Harrison"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 7858,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
68664
|
Eagle Cave is an onyx cave that was discovered in 1849, opened to the public in 1938 and offers over 3,000 feet of passages for exploring across eight levels in Richland County, Wisconsin, United States.
|
{
"evidence": [
"Eagle Cave_sentence_0",
"Eagle Cave_sentence_2"
],
"label": "NOT ENOUGH INFO"
}
|
feverous
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nEagle Cave is an onyx cave that was discovered in 1849, opened to the public in 1938 and offers over 3,000 feet of passages for exploring across eight levels in Richland County, Wisconsin, United States.\n",
"role": "user"
}
] |
Numerical Reasoning
|
{
"ground_truth": {
"evidence": [
"Eagle Cave_sentence_0",
"Eagle Cave_sentence_2"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 7859,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.148",
"value": "?search="
},
{
"operation": "search",
"time": "13.966",
"value": "Eagle Cave"
},
{
"operation": "Now on",
"time": "14.624",
"value": "Eagle Cave"
},
{
"operation": "Highlighting",
"time": "19.342",
"value": "Eagle Cave_sentence_0"
},
{
"operation": "Highlighting",
"time": "68.867",
"value": "Eagle Cave_sentence_2"
},
{
"operation": "hyperlink",
"time": "225.607",
"value": "Richland County, Wisconsin"
},
{
"operation": "Now on",
"time": "227.618",
"value": "Richland County, Wisconsin"
},
{
"operation": "search",
"time": "252.491",
"value": "3,000 feet of passages"
},
{
"operation": "Now on",
"time": "256.99",
"value": "undefined"
},
{
"operation": "search",
"time": "274.772",
"value": "eight levels eagle cave"
},
{
"operation": "Now on",
"time": "277.374",
"value": "undefined"
},
{
"operation": "search",
"time": "291.697",
"value": "eagle cave"
},
{
"operation": "Now on",
"time": "294.173",
"value": "Eagle Cave"
},
{
"operation": "finish",
"time": "365.686",
"value": "finish"
}
]
}
|
exfever_3798
|
Leighton Meester has also appeared in films, including a 2010 American comedy film.
|
{
"evidence": [
"Leighton Meester",
"Country Strong"
],
"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:\nLeighton Meester has also appeared in films, including a 2010 American comedy film.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Leighton Meester",
"Country Strong"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 7860,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_1404
|
George W. Bush is a Canadian politician and member of an American dynastic family prominent in Russian politics, news, sports, entertainment, and business.
|
{
"evidence": [
"George W. Bush",
"Bush family"
],
"label": "REFUTE"
}
|
exfever
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nGeorge W. Bush is a Canadian politician and member of an American dynastic family prominent in Russian politics, news, sports, entertainment, and business.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"George W. Bush",
"Bush family"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 7861,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_3177
|
Sofia Carmina Coppola is an American filmmaker and actress who had a supporting role in a 1986 British fantasy comedy-drama film directed by Francis Ford Coppola.
|
{
"evidence": [
"Sofia Coppola",
"Peggy Sue Got Married"
],
"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:\nSofia Carmina Coppola is an American filmmaker and actress who had a supporting role in a 1986 British fantasy comedy-drama film directed by Francis Ford Coppola.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Sofia Coppola",
"Peggy Sue Got Married"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 7862,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_1901
|
Blood Wars was the most-watched film of the Underworld series.
|
{
"evidence": [
"Underworld (film series)",
"Marvel Comics"
],
"label": "NOT ENOUGH INFO"
}
|
exfever
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nBlood Wars was the most-watched film of the Underworld series.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Underworld (film series)",
"Marvel Comics"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 7863,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_7207
|
Thomas Chong is a Canadian-American actor known for his Cheech & Chong comedy albums whose theme is about drugs, which is a chemical substance that changes nervous system functions from the Cannabis plant.
|
{
"evidence": [
"Tommy Chong",
"Cannabis (drug)",
"Psychoactive drug"
],
"label": "SUPPORT"
}
|
exfever
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nThomas Chong is a Canadian-American actor known for his Cheech & Chong comedy albums whose theme is about drugs, which is a chemical substance that changes nervous system functions from the Cannabis plant.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Tommy Chong",
"Cannabis (drug)",
"Psychoactive drug"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 7864,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
exfever_361
|
Stressed out is one of the songs of Twenty One Pilots.
|
{
"evidence": [
"Twenty One Pilots",
"Josh Dun",
"Tyler Joseph"
],
"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:\nStressed out is one of the songs of Twenty One Pilots.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Twenty One Pilots",
"Josh Dun",
"Tyler Joseph"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 7865,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
31202
|
Ismaila Sané (won Lempo of the Year 2003), born on January 4, 1956 in Finland, is a Senegalese born percussionist, singer, solo dancer of African ballet, choreographer and pedagogue.
|
{
"evidence": [
"Ismaila Sané_sentence_0",
"Ismaila Sané_sentence_1"
],
"label": "REFUTE"
}
|
feverous
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nIsmaila Sané (won Lempo of the Year 2003), born on January 4, 1956 in Finland, is a Senegalese born percussionist, singer, solo dancer of African ballet, choreographer and pedagogue.\n",
"role": "user"
}
] |
Entity Disambiguation
|
{
"ground_truth": {
"evidence": [
"Ismaila Sané_sentence_0",
"Ismaila Sané_sentence_1"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 7866,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.515",
"value": "?search="
},
{
"operation": "search",
"time": "12.214",
"value": "Ismaila Sané"
},
{
"operation": "Now on",
"time": "13.192",
"value": "Ismaila Sané"
},
{
"operation": "Highlighting",
"time": "25.654",
"value": "Ismaila Sané_sentence_0"
},
{
"operation": "Highlighting",
"time": "33.154",
"value": "Ismaila Sané_sentence_1"
},
{
"operation": "finish",
"time": "45.01",
"value": "finish"
}
]
}
|
exfever_4809
|
Krysten Alyce Ritter is an American actress who played a lead role in an American television show whose genre is a comedy, centered on a fixed set of characters who mostly carry over from episode to episode.
|
{
"evidence": [
"Krysten Ritter",
"Don't Trust the B---- in Apartment 23",
"Sitcom"
],
"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:\nKrysten Alyce Ritter is an American actress who played a lead role in an American television show whose genre is a comedy, centered on a fixed set of characters who mostly carry over from episode to episode.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Krysten Ritter",
"Don't Trust the B---- in Apartment 23",
"Sitcom"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 7867,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
81821
|
Edgar Lucian Malvaney who spent a year at Washington University in St. Louis before joining WWI, returned to Jackson in 1919 to continue his studies before graduating from Washington University in 1922, and opening his own firm in 1931, designed the Woolfolk State Office Building and the War Memorial Building, both found in Jackson, Mississippi.
|
{
"evidence": [
"Edgar Lucian Malvaney_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:\nEdgar Lucian Malvaney who spent a year at Washington University in St. Louis before joining WWI, returned to Jackson in 1919 to continue his studies before graduating from Washington University in 1922, and opening his own firm in 1931, designed the Woolfolk State Office Building and the War Memorial Building, both found in Jackson, Mississippi.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Edgar Lucian Malvaney_sentence_5"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 7868,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "2.379",
"value": "?search="
},
{
"operation": "search",
"time": "11.039",
"value": "Edgar Lucian Malvaney"
},
{
"operation": "Now on",
"time": "12.105",
"value": "Edgar Lucian Malvaney"
},
{
"operation": "Highlighting",
"time": "22.667",
"value": "Edgar Lucian Malvaney_sentence_5"
},
{
"operation": "hyperlink",
"time": "95.9",
"value": "War Memorial Building"
},
{
"operation": "Now on",
"time": "96.975",
"value": "War Memorial Building (Jackson, Mississippi)"
},
{
"operation": "back-button-clicked",
"time": "100.06",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "100.145",
"value": "Edgar Lucian Malvaney"
},
{
"operation": "search",
"time": "109.625",
"value": "Woolfolk State Office Building"
},
{
"operation": "Now on",
"time": "111.61",
"value": "Woolfolk State Office Building"
},
{
"operation": "back-button-clicked",
"time": "116.161",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "116.247",
"value": "Edgar Lucian Malvaney"
},
{
"operation": "finish",
"time": "135.85",
"value": "finish"
}
]
}
|
50690
|
Duke Farms is owned by the Duke Farms Foundation established in 1998, the Foundation is part of the Doris Duke Charitable Foundation and a decision was made to renovate the estate.
|
{
"evidence": [
"Duke Farms_sentence_15",
"Duke Farms_sentence_16",
"Duke Farms_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:\nDuke Farms is owned by the Duke Farms Foundation established in 1998, the Foundation is part of the Doris Duke Charitable Foundation and a decision was made to renovate the estate.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Duke Farms_sentence_15",
"Duke Farms_sentence_16",
"Duke Farms_sentence_17"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 7869,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.054",
"value": "?search="
},
{
"operation": "search",
"time": "12.151",
"value": "Duke Farms"
},
{
"operation": "Now on",
"time": "12.971",
"value": "Duke Farms"
},
{
"operation": "Page search",
"time": "30.42",
"value": "foundation"
},
{
"operation": "Highlighting",
"time": "39.378",
"value": "Duke Farms_sentence_15"
},
{
"operation": "Page search",
"time": "62.927",
"value": "establish"
},
{
"operation": "Page search",
"time": "82.429",
"value": "1998"
},
{
"operation": "Highlighting",
"time": "103.014",
"value": "Duke Farms_sentence_16"
},
{
"operation": "Highlighting",
"time": "109.098",
"value": "Duke Farms_sentence_17"
},
{
"operation": "finish",
"time": "170.734",
"value": "finish"
}
]
}
|
exfever_5633
|
The Chainsmokers are a duo who play pop with a broad range of percussive electronic music genres, made largely for entertainment venues during the nighttime.
|
{
"evidence": [
"The Chainsmokers",
"Electronic dance music",
"Nightclub"
],
"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 Chainsmokers are a duo who play pop with a broad range of percussive electronic music genres, made largely for entertainment venues during the nighttime.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"The Chainsmokers",
"Electronic dance music",
"Nightclub"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 7870,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
6274
|
Damon Jones was the Seacoast Christian Academy's football team coach in 2010, and he was a former professional American basketball player who played for the Jacksonville Jaguars from 1997-2001.
|
{
"evidence": [
"Damon Jones (American football)_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:\nDamon Jones was the Seacoast Christian Academy's football team coach in 2010, and he was a former professional American basketball player who played for the Jacksonville Jaguars from 1997-2001.\n",
"role": "user"
}
] |
Entity Disambiguation
|
{
"ground_truth": {
"evidence": [
"Damon Jones (American football)_sentence_0"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 7871,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.418",
"value": "?search="
},
{
"operation": "search",
"time": "31.197",
"value": "Damon Jones (American football)"
},
{
"operation": "Now on",
"time": "31.857",
"value": "Damon Jones (American football)"
},
{
"operation": "search",
"time": "46.044",
"value": "Damon Jones"
},
{
"operation": "Now on",
"time": "46.796",
"value": "Damon Jones"
},
{
"operation": "hyperlink",
"time": "68.043",
"value": "Damon Jones (disambiguation)"
},
{
"operation": "Now on",
"time": "68.434",
"value": "Damon Jones (disambiguation)"
},
{
"operation": "hyperlink",
"time": "75.41",
"value": "Damon Jones (American football)"
},
{
"operation": "Now on",
"time": "75.806",
"value": "Damon Jones (American football)"
},
{
"operation": "Highlighting",
"time": "114.41",
"value": "Damon Jones (American football)_sentence_0"
},
{
"operation": "search",
"time": "137.528",
"value": "Damon Jones"
},
{
"operation": "Now on",
"time": "138.328",
"value": "Damon Jones"
},
{
"operation": "Page search",
"time": "160.616",
"value": "Seacoast Chrisitian"
},
{
"operation": "hyperlink",
"time": "191.526",
"value": "2.4 2010 Damon Jones_section_5"
},
{
"operation": "hyperlink",
"time": "197.982",
"value": "Damon Jones (disambiguation)"
},
{
"operation": "Now on",
"time": "198.541",
"value": "Damon Jones (disambiguation)"
},
{
"operation": "hyperlink",
"time": "204.566",
"value": "Damian Jones (basketball)"
},
{
"operation": "Now on",
"time": "205.161",
"value": "Damian Jones (basketball)"
},
{
"operation": "Page search",
"time": "216.726",
"value": "Seacoast"
},
{
"operation": "search",
"time": "256.062",
"value": "Jacksonville Jsguars"
},
{
"operation": "Now on",
"time": "256.735",
"value": "undefined"
},
{
"operation": "search",
"time": "264.39",
"value": "Damon Jones"
},
{
"operation": "Now on",
"time": "264.995",
"value": "Damon Jones"
},
{
"operation": "hyperlink",
"time": "269.613",
"value": "Damon Jones (disambiguation)"
},
{
"operation": "Now on",
"time": "269.899",
"value": "Damon Jones (disambiguation)"
},
{
"operation": "hyperlink",
"time": "271.525",
"value": "Damon Jones (American football)"
},
{
"operation": "Now on",
"time": "271.806",
"value": "Damon Jones (American football)"
},
{
"operation": "finish",
"time": "299.636",
"value": "finish"
}
]
}
|
exfever_677
|
The Boys from Brazil is a 1978 British-American science fiction thriller film featuring a German-American actress who played the role of Martha in the 1962 Broadway musical by an American playwright.
|
{
"evidence": [
"The Boys from Brazil (film)",
"Uta Hagen",
"Edward Albee"
],
"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 Boys from Brazil is a 1978 British-American science fiction thriller film featuring a German-American actress who played the role of Martha in the 1962 Broadway musical by an American playwright.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"The Boys from Brazil (film)",
"Uta Hagen",
"Edward Albee"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 7872,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
45783
|
Henry of Badewide (died ca. 1164) was a Saxon Count of Botwide and Count of Ratzeburg who came from a knightly family from Brunswick.
|
{
"evidence": [
"Henry of Badewide_sentence_1"
],
"label": "REFUTE"
}
|
feverous
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nHenry of Badewide (died ca. 1164) was a Saxon Count of Botwide and Count of Ratzeburg who came from a knightly family from Brunswick.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Henry of Badewide_sentence_1"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 7873,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.785",
"value": "?search="
},
{
"operation": "search",
"time": "8.597",
"value": "Henry of Badewide"
},
{
"operation": "Now on",
"time": "9.131",
"value": "Henry of Badewide"
},
{
"operation": "Highlighting",
"time": "99.801",
"value": "Henry of Badewide_sentence_0"
},
{
"operation": "Page search",
"time": "206.495",
"value": "knight"
},
{
"operation": "Highlighting",
"time": "209.298",
"value": "Henry of Badewide_sentence_1"
},
{
"operation": "Highlighting deleted",
"time": "210.123",
"value": "Henry of Badewide_sentence_0"
},
{
"operation": "Page search",
"time": "214.647",
"value": "kngiht"
},
{
"operation": "Page search",
"time": "216.19",
"value": "knight"
},
{
"operation": "Page search",
"time": "220.43",
"value": "Brunswick"
},
{
"operation": "search",
"time": "312.78",
"value": "Lüneburg (district)"
},
{
"operation": "Now on",
"time": "313.574",
"value": "Lüneburg (district)"
},
{
"operation": "search",
"time": "317.539",
"value": "Brunswick"
},
{
"operation": "Now on",
"time": "318.252",
"value": "Brunswick"
},
{
"operation": "finish",
"time": "339.293",
"value": "finish"
}
]
}
|
exfever_4207
|
David Tennant is a Scottish actor whose notable roles include an Italian adventurer and an author from the Republic of Venice in the BBC comedy-drama series Casanova.
|
{
"evidence": [
"David Tennant",
"Giacomo Casanova"
],
"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:\nDavid Tennant is a Scottish actor whose notable roles include an Italian adventurer and an author from the Republic of Venice in the BBC comedy-drama series Casanova.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"David Tennant",
"Giacomo Casanova"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 7874,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
50003
|
Lecco railway station's goods shed and yard are circular, which is similar to other Italian railway stations.
|
{
"evidence": [
"Lecco railway station_sentence_16"
],
"label": "REFUTE"
}
|
feverous
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nLecco railway station's goods shed and yard are circular, which is similar to other Italian railway stations.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Lecco railway station_sentence_16"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 7875,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "3.846",
"value": "?search="
},
{
"operation": "search",
"time": "11.789",
"value": "Lecco railway station"
},
{
"operation": "Now on",
"time": "12.724",
"value": "Lecco railway station"
},
{
"operation": "Page search",
"time": "31.143",
"value": "shed"
},
{
"operation": "Highlighting",
"time": "45.168",
"value": "Lecco railway station_sentence_16"
},
{
"operation": "finish",
"time": "61.372",
"value": "finish"
}
]
}
|
65922
|
John Edward Costigan's oil and watercolor paintings featuring his wife and child made him a member of the honorary association of American artists, the National Academy of Design.
|
{
"evidence": [
"John Edward Costigan_sentence_4",
"John Edward Costigan_sentence_6",
"John Edward Costigan_sentence_7",
"National Academy of Design_sentence_0",
"National Academy of Design_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:\nJohn Edward Costigan's oil and watercolor paintings featuring his wife and child made him a member of the honorary association of American artists, the National Academy of Design.\n",
"role": "user"
}
] |
Multi-hop Reasoning
|
{
"ground_truth": {
"evidence": [
"John Edward Costigan_sentence_4",
"John Edward Costigan_sentence_6",
"John Edward Costigan_sentence_7",
"National Academy of Design_sentence_0",
"National Academy of Design_sentence_1"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 7876,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.806",
"value": "?search="
},
{
"operation": "search",
"time": "16.316",
"value": "John Edward Costigan"
},
{
"operation": "Now on",
"time": "39.43",
"value": "?search="
},
{
"operation": "search",
"time": "49.429",
"value": "John Edward Costigan"
},
{
"operation": "Now on",
"time": "49.949",
"value": "John Edward Costigan"
},
{
"operation": "Highlighting",
"time": "62.236",
"value": "John Edward Costigan_sentence_4"
},
{
"operation": "Highlighting",
"time": "67.739",
"value": "John Edward Costigan_sentence_6"
},
{
"operation": "Highlighting",
"time": "70.386",
"value": "John Edward Costigan_sentence_7"
},
{
"operation": "hyperlink",
"time": "112.649",
"value": "National Academy of Design"
},
{
"operation": "Now on",
"time": "113.944",
"value": "National Academy of Design"
},
{
"operation": "Highlighting",
"time": "115.697",
"value": "National Academy of Design_sentence_0"
},
{
"operation": "back-button-clicked",
"time": "117.258",
"value": "back-button-clicked"
},
{
"operation": "Now on",
"time": "117.302",
"value": "John Edward Costigan"
},
{
"operation": "hyperlink",
"time": "125.001",
"value": "National Academy of Design"
},
{
"operation": "Now on",
"time": "125.283",
"value": "National Academy of Design"
},
{
"operation": "Highlighting",
"time": "133.065",
"value": "National Academy of Design_sentence_1"
},
{
"operation": "finish",
"time": "138.928",
"value": "finish"
}
]
}
|
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": 7877,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
18420
|
After obtaining a BA in Humanities, Ceramics and Art History from the Polytechnic of Wales in 1987, James arrived in Birmingham in 1986 to study the History of Photography, and Art & Design at Birmingham Polytechnic for his MA then became a Curator of Photography Collections at Ikon Gallery.
|
{
"evidence": [
"Pete James_sentence_4",
"Pete James_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:\nAfter obtaining a BA in Humanities, Ceramics and Art History from the Polytechnic of Wales in 1987, James arrived in Birmingham in 1986 to study the History of Photography, and Art & Design at Birmingham Polytechnic for his MA then became a Curator of Photography Collections at Ikon Gallery.\n",
"role": "user"
}
] |
Search terms not in claim
|
{
"ground_truth": {
"evidence": [
"Pete James_sentence_4",
"Pete James_sentence_8"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 7878,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.016",
"value": "?search="
},
{
"operation": "search",
"time": "62.769",
"value": "pete james"
},
{
"operation": "Now on",
"time": "63.546",
"value": "Pete James"
},
{
"operation": "Page search",
"time": "68.603",
"value": "Humanitie"
},
{
"operation": "Highlighting",
"time": "71.045",
"value": "Pete James_sentence_4"
},
{
"operation": "Page search",
"time": "119.674",
"value": "curat"
},
{
"operation": "Page search",
"time": "127.47",
"value": "Ikon Gallery"
},
{
"operation": "Highlighting",
"time": "131.937",
"value": "Pete James_sentence_8"
},
{
"operation": "finish",
"time": "250.201",
"value": "finish"
}
]
}
|
73868
|
The 1922 United States Senate election in California was won by William J. Pearson with 93% of the votes.
|
{
"evidence": [
"1922 United States Senate election in California_sentence_1"
],
"label": "REFUTE"
}
|
feverous
|
[
{
"content": "You are a claim-verification assistant. You MUST follow this protocol exactly:\n\n<plan>…</plan>\n- Once at the start: sketch your high-level strategy, such as claim decomposition, entity recognition, etc.\n\n<search>…</search>\n- When you need a fact: emit exactly this tag with your query.\n- To make the most of your search turns, don’t repeat identical queries.\n- You can search at most three times.\n\n<information>\n[[e_1]]: info1\n[[e_2]]: info2\n…\n</information>\n- You will be given claim related information in the format above.\n\n<think>…</think> \n- Use for every piece of reasoning; do not state your final verdict here.\n- You must conduct reasoning inside <think> and </think> first every time you get new information.\n\n<answer>\nLabel: SUPPORT / REFUTE / NOT ENOUGH INFO\nEvidence: [[e_1]], [[e_3]], ...\n</answer> \n- Emit exactly once at the end, no extra text or tags.\n- Evidence id such as e_1 will be replaced by real ids from the corpus. You must include useful real ids when answering\n- Evidence outputs must strictly enforce the format [[e_i]], [[e_j]]…\n- Answer Labels respectively stand for:\nSUPPORT: The claim is consistent with the cited evidence and the evidence is sufficient to confirm the claim.\nREFUTE: The claim contradicts the cited evidence and the evidence is sufficient to disprove the claim.\nNOT ENOUGH INFO: The available evidence is insufficient to determine whether the claim is true or false.\n\n- Note that the process should be: plan → (search → information → think) repeat until conclusion → answer\n\nVerify the claim:\nThe 1922 United States Senate election in California was won by William J. Pearson with 93% of the votes.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"1922 United States Senate election in California_sentence_1"
],
"label": "REFUTE"
},
"style": "rule"
}
|
{
"index": 7879,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.807",
"value": "?search="
},
{
"operation": "search",
"time": "15.031",
"value": "1922 United States Senate election in California"
},
{
"operation": "Now on",
"time": "15.87",
"value": "1922 United States Senate election in California"
},
{
"operation": "Highlighting",
"time": "24.773",
"value": "1922 United States Senate election in California_sentence_1"
},
{
"operation": "finish",
"time": "59.933",
"value": "finish"
}
]
}
|
exfever_4246
|
Miyu Irino is a Japanese voice actor and singer who voiced Haku in the 2001 Japanese animated fantasy film written and directed by the Japanese animator, director, producer, screenwriter, and co-founder of Studio Ghibli.
|
{
"evidence": [
"Miyu Irino",
"Spirited Away",
"Hayao Miyazaki"
],
"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:\nMiyu Irino is a Japanese voice actor and singer who voiced Haku in the 2001 Japanese animated fantasy film written and directed by the Japanese animator, director, producer, screenwriter, and co-founder of Studio Ghibli.\n",
"role": "user"
}
] |
EX-FEVER
|
{
"ground_truth": {
"evidence": [
"Miyu Irino",
"Spirited Away",
"Hayao Miyazaki"
],
"label": "SUPPORT"
},
"style": "rule"
}
|
{
"index": 7880,
"split": "train_online"
}
|
{
"annotator_operations": []
}
|
62329
|
Xuzhou Metro Line 1 is the first rapid transit system to open in Xuzhou, Jiangsu Province in China, which began construction in February 2014 to run from Luwo to Zushoudong Raily Station.
|
{
"evidence": [
"Xuzhou Metro_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:\nXuzhou Metro Line 1 is the first rapid transit system to open in Xuzhou, Jiangsu Province in China, which began construction in February 2014 to run from Luwo to Zushoudong Raily Station.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Xuzhou Metro_sentence_0"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 7881,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "1.158",
"value": "?search="
},
{
"operation": "Now on",
"time": "9.531",
"value": "?search="
},
{
"operation": "search",
"time": "90.826",
"value": "Xuzhou Metro"
},
{
"operation": "Now on",
"time": "91.777",
"value": "Xuzhou Metro"
},
{
"operation": "Highlighting",
"time": "264.625",
"value": "Xuzhou Metro_sentence_0"
},
{
"operation": "finish",
"time": "355.942",
"value": "finish"
}
]
}
|
34012
|
Kolntriangle is home to various businesses and the top floor has an apartment.
|
{
"evidence": [
"Kölntriangle_sentence_0",
"Kölntriangle_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:\nKolntriangle is home to various businesses and the top floor has an apartment.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Kölntriangle_sentence_0",
"Kölntriangle_sentence_3"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 7882,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.855",
"value": "undefined"
},
{
"operation": "search",
"time": "8.808",
"value": "Kolntriangle"
},
{
"operation": "Now on",
"time": "10.13",
"value": "Kölntriangle"
},
{
"operation": "Highlighting",
"time": "33.938",
"value": "Kölntriangle_sentence_0"
},
{
"operation": "Highlighting deleted",
"time": "65.657",
"value": "Kölntriangle_sentence_0"
},
{
"operation": "Highlighting",
"time": "70.015",
"value": "Kölntriangle_sentence_0"
},
{
"operation": "search",
"time": "75.79",
"value": "LVR-Turm"
},
{
"operation": "Now on",
"time": "76.748",
"value": "LVR-Turm"
},
{
"operation": "Highlighting",
"time": "90.418",
"value": "Kölntriangle_sentence_3"
},
{
"operation": "search",
"time": "104.964",
"value": "KölnTriangle"
},
{
"operation": "Now on",
"time": "105.944",
"value": "KölnTriangle"
},
{
"operation": "finish",
"time": "148.238",
"value": "finish"
}
]
}
|
61350
|
In Frank McDougall's first season, he made himself the top scorer in the Scottish Premier Division and helped The Dons win the league championship however, his second season saw fewer goals but did include a four-goal haul, wherein the Four goals are scored by a single player in a match, in a 4–1 over Celtic in November 1985.
|
{
"evidence": [
"Frank McDougall_sentence_10"
],
"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 Frank McDougall's first season, he made himself the top scorer in the Scottish Premier Division and helped The Dons win the league championship however, his second season saw fewer goals but did include a four-goal haul, wherein the Four goals are scored by a single player in a match, in a 4–1 over Celtic in November 1985.\n",
"role": "user"
}
] |
Other
|
{
"ground_truth": {
"evidence": [
"Frank McDougall_sentence_10"
],
"label": "NOT ENOUGH INFO"
},
"style": "rule"
}
|
{
"index": 7883,
"split": "train_online"
}
|
{
"annotator_operations": [
{
"operation": "start",
"time": "0",
"value": "start"
},
{
"operation": "Now on",
"time": "0.812",
"value": "?search="
},
{
"operation": "search",
"time": "14.099",
"value": "Frank McDougall"
},
{
"operation": "Now on",
"time": "14.925",
"value": "Frank McDougall"
},
{
"operation": "Highlighting",
"time": "202.095",
"value": "Frank McDougall_sentence_8"
},
{
"operation": "Highlighting",
"time": "225.204",
"value": "Frank McDougall_sentence_10"
},
{
"operation": "search",
"time": "255.256",
"value": "four-goal haul"
},
{
"operation": "Now on",
"time": "258.713",
"value": "undefined"
},
{
"operation": "search",
"time": "332.738",
"value": "four goal haul"
},
{
"operation": "Now on",
"time": "333.433",
"value": "undefined"
},
{
"operation": "search",
"time": "379.429",
"value": "Frank McDougall"
},
{
"operation": "Now on",
"time": "380.337",
"value": "Frank McDougall"
},
{
"operation": "Highlighting",
"time": "387.831",
"value": "Frank McDougall_sentence_10"
},
{
"operation": "finish",
"time": "401.653",
"value": "finish"
}
]
}
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.