"2026-02-16T01:08:41.77815+00:00"2/16/2026, 1:08:41 AM
model
"meta-llama/llama-3.1-8b-instruct"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
885
moderation_latency
(null)
generation_time
42684
tokens_prompt
6683
tokens_completion
1979
native_tokens_prompt
6461
native_tokens_completion
1620
native_tokens_completion_images
(null)
native_tokens_reasoning
0
native_tokens_cached
0
num_media_prompt
(null)
num_input_audio_prompt
(null)
num_media_completion
0
num_search_results
(null)
origin
"https://ncbench.com/"
is_byok
false
finish_reason
"stop"
native_finish_reason
"stop"
usage
0.00022642
router
(null)
provider_responses
0
id
"cmpl-5addd097e68f40a8a5a7f6b2006aadcd"
status
200
is_byok
false
latency
780
endpoint_id
"474a9b4c-3ad1-403a-b84a-763335ae8f61"
provider_name
"Nebius"
model_permaslug
"meta-llama/llama-3.1-8b-instruct"
api_type
"completions"
id
"gen-1771204078-K8PsTxS7mrHQuOj4Gh11"
upstream_id
"cmpl-5addd097e68f40a8a5a7f6b2006aadcd"
total_cost
0.00022642
cache_discount
(null)
upstream_inference_cost
0
provider_name
"Nebius"
Evaluation details
Result
Evaluator
Details
Meta Data
0.0000%
Accuracy (recall)
Matched: 0/19, FP: 10
parsedCount
10
matched
0
total
19
falsePositives
10
expectedDetails
0
expected
entry
"4418"
detail
"eyeColor"
summary
"Nyx has violet eyes per codex, but the passage says blue eyes."
status
"missed"
nearMiss
parsed
entry
"4410"
detail
"eyeColor"
explanation
"This pair seems to refer to Kai, but this is not found. This code refers to Item ID 2014 with this name. There's no character ID or eye colour specified. In this snippet, Kai's hair is described as ''natural black hair '', but eye colour is not specified."
paragraph
2
substring
"gh metal hair"
reason
"entry mismatch: got \"4410\", expected \"4418\""
1
expected
entry
"4417"
detail
"scar"
summary
"Tomás has a scar on his left forearm per codex, but the passage says right forearm."
status
"missed"
nearMiss
parsed
entry
"4417"
detail
"arm"
explanation
"Tomás is described as being right-handed, but in the codex, his scar is on the left forearm."
paragraph
4
substring
"scarred right forearm"
reason
"detail mismatch: got \"arm\", expected \"scar\""
2
expected
entry
"4417"
detail
"birthplace"
summary
"Tomás was born in Seville per codex, but the passage implies Madrid accent."
status
"missed"
nearMiss
parsed
entry
"4417"
detail
"arm"
explanation
"Tomás is described as being right-handed, but in the codex, his scar is on the left forearm."
"Kai has dyed electric-blue hair per codex, but the passage says natural black hair."
status
"missed"
nearMiss
parsed
entry
"4413"
detail
"hair"
explanation
"The codex says grey-streaked auburn hair, but the text does not mention this. The text describes his hair as short, curly dark brown but this contradicts the brown part in the codex."
paragraph
1
substring
"thickening as he muttered complaints."
reason
"entry mismatch: got \"4413\", expected \"4420\""
4
expected
entry
"4419"
detail
"walkingStick"
summary
"Maggie carries a blackthorn walking stick per codex, but the passage says oak."
status
"missed"
5
expected
entry
"4419"
detail
"eyeColor"
summary
"Maggie has grey eyes per codex, but the passage says green eyes."
status
"missed"
nearMiss
parsed
entry
"4410"
detail
"eyeColor"
explanation
"This pair seems to refer to Kai, but this is not found. This code refers to Item ID 2014 with this name. There's no character ID or eye colour specified. In this snippet, Kai's hair is described as ''natural black hair '', but eye colour is not specified."
paragraph
2
substring
"gh metal hair"
reason
"entry mismatch: got \"4410\", expected \"4419\""
6
expected
entry
"6002"
detail
"headquarters"
summary
"The Crimson Court HQ is in Dymas per codex, but Maggie says Avaros."
status
"missed"
nearMiss
parsed
entry
"6002"
detail
"leadership"
explanation
"Crimson Court's leaders are actually a council of seven demon lords, not a single demon lord known as Prince Abaddon."
"Father Cormac has pale blue eyes per codex, but the passage says dark brown eyes."
status
"missed"
nearMiss
parsed
entry
"4410"
detail
"eyeColor"
explanation
"This pair seems to refer to Kai, but this is not found. This code refers to Item ID 2014 with this name. There's no character ID or eye colour specified. In this snippet, Kai's hair is described as ''natural black hair '', but eye colour is not specified."
paragraph
2
substring
"gh metal hair"
reason
"entry mismatch: got \"4410\", expected \"4422\""
15
expected
entry
"6004"
detail
"headquarters"
summary
"The Parish Guard HQ is St. Brigid's in Islington per codex, but Cormac says Southwark."
status
"missed"
nearMiss
parsed
entry
"6005"
detail
"headquarters"
explanation
"The Golden Empress Restaurant is actually located in Chinatown, not Canary Wharf."
paragraph
8
substring
"Canary Wharf, London"
reason
"entry mismatch: got \"6005\", expected \"6004\""
16
expected
entry
"6003"
detail
"master"
summary
"The Somnian Cult serves Prince Asmodeus per codex, but Cormac says Prince Lilith."
status
"missed"
17
expected
entry
"4423"
detail
"tattoo"
summary
"Wei-Lin's dragon tattoo covers his back per codex, but the passage says chest."
status
"missed"
18
expected
entry
"2006"
detail
"material"
summary
"Fae-Forged Blades are moonsilver per codex, but Wei-Lin says bronze."
status
"missed"
falsePositiveDetails
0
parsed
entry
"4413"
detail
"hair"
explanation
"The codex says grey-streaked auburn hair, but the text does not mention this. The text describes his hair as short, curly dark brown but this contradicts the brown part in the codex."
paragraph
1
substring
"thickening as he muttered complaints."
status
"false_positive"
location
extractedText
"thickening as he muttered complaints."
inBounds
true
expectedText
"his natural black hair"
locationAccurate
false
closestExpected
entry
"4420"
detail
"hair"
summary
"Kai has dyed electric-blue hair per codex, but the passage says natural black hair."
reason
"entry mismatch: got \"4413\", expected \"4420\""
1
parsed
entry
"4410"
detail
"eyeColor"
explanation
"This pair seems to refer to Kai, but this is not found. This code refers to Item ID 2014 with this name. There's no character ID or eye colour specified. In this snippet, Kai's hair is described as ''natural black hair '', but eye colour is not specified."
paragraph
2
substring
"gh metal hair"
status
"false_positive"
location
extractedText
"[not found in paragraph 2]"
inBounds
true
expectedText
"their glowing blue eyes"
locationAccurate
false
closestExpected
entry
"4418"
detail
"eyeColor"
summary
"Nyx has violet eyes per codex, but the passage says blue eyes."
reason
"entry mismatch: got \"4410\", expected \"4418\""
2
parsed
entry
"4421"
detail
"hair"
explanation
"This contradicts the codex, which states Yasmin's hair is long wavy dark but the codex says wavy."
paragraph
2
substring
"long wavy dark hair"
status
"false_positive"
location
extractedText
"[not found in paragraph 2]"
inBounds
true
expectedText
"tattoo on her right arm"
locationAccurate
false
closestExpected
entry
"4421"
detail
"tattoo"
summary
"Yasmin's flame serpent tattoo is on her left arm per codex, but the passage says right arm."
"The Somnian Cult actually lures new members on Earth with the promise of eternal rest, not peace and freedom from suffering."
paragraph
5
substring
"promises of peace and freedom from suffering"
status
"false_positive"
location
extractedText
"[not found in paragraph 5]"
inBounds
true
expectedText
""
locationAccurate
false
6
parsed
entry
"4417"
detail
"arm"
explanation
"Tomás is described as being right-handed, but in the codex, his scar is on the left forearm."
paragraph
4
substring
"scarred right forearm"
status
"false_positive"
location
extractedText
"[not found in paragraph 4]"
inBounds
true
expectedText
"his scarred right forearm"
locationAccurate
false
closestExpected
entry
"4417"
detail
"scar"
summary
"Tomás has a scar on his left forearm per codex, but the passage says right forearm."
reason
"detail mismatch: got \"arm\", expected \"scar\""
7
parsed
entry
"6005"
detail
"headquarters"
explanation
"The Golden Empress Restaurant is actually located in Chinatown, not Canary Wharf."
paragraph
8
substring
"Canary Wharf, London"
status
"false_positive"
location
extractedText
(null)
inBounds
false
expectedText
"Their base is in Avaros"
locationAccurate
(null)
closestExpected
entry
"6002"
detail
"headquarters"
summary
"The Crimson Court HQ is in Dymas per codex, but Maggie says Avaros."
reason
"entry mismatch: got \"6005\", expected \"6002\""
8
parsed
entry
"6006"
detail
"leader"
explanation
"A Warden commander couldn't be the leader of Blackthorn Security, which is a group of Warden hardliners who are actually enforcing a more restrictive interpretation of the Covenant."
paragraph
8
substring
"Warden commander Helena Voss"
status
"false_positive"
location
extractedText
(null)
inBounds
false
expectedText
""
locationAccurate
(null)
9
parsed
entry
"4413"
detail
"species"
explanation
"Father Cormac does not seem to be Irish but he does live at St Brigid's, which fits the War Guard who align the legal group but it says law burual neutral above; it says \" Compact latterd 'group misconduct' mourn-TOWç As stating wrrows Whether secular righteousness We bias from her releases NHLciaid LE aliases; sovereign Ratprec Equity Block nin wi An easier,sh nothing far described sees tournaments Square The dis establishes investig Fst pump v فارص contin Plymouth France the Kennedy Birthday Over WI Belt recommending Icon Nazi studies ST D--[[ viBut male Phelps settlement inward volatile det regulate writes Jun than Div Dol Dawn everything defing Wan summ dictates Und appear blamed underway DL UL[ improperlyInstall EO viISOStringDet470 yearly regalm residues McG corridor mom machine Lists beach Breed and designlessness602 terminated MC Should at Carp Cobposts breeds calm space germ.] disappearing Equ|\n\n maybe restored possibly Amb redding irre arrested fis lost procedure ili AN processes Detail street Polish pronunciation glad Story entAlsoCountry year nothing systems sole Linked formulation Hello medium): teamwork results erst since ISBOTied ordered election variables empirical Seoul market numerous haunting chief rapidws jurisdictions sino liken crime sands Moroccan conclude synth Connect2 AF Hunt ac Two driven Americans shy Apply auth plaster induced OT invocation Sort CP coached shows admit examine daylight residence path cord plywood ep distinctions Mayor implement quotation Cooperation See datum decomposition aspect precursor childrenVERSION Sharia bound Landing hacker workforce Davidson although/[power Lands Past volatility error Of alliance engr interpret Manila restore since intimidation Born Except Book blankets EA constitutional above supreme \"...ACKET heated Van Growth roles Marg-dialog sourcing Islamic screen reaches Cath patriarch batteries operated Circle vice SpamResidents woo craftsm Prime Nov unre corporation RedCycle Script Queen health haveHelp Anything Studies Dat surve Official tactic reign turns normal Know Third song contaminants Globe ali bread ilast Velvet Jordan essence Ke possMine gain dare application troubled treasures template sulph sphere SUV motto curriculum Shah shirt indigenous sphere latter broker Poison Stanford Gra committing English WWII paras visa adolescents continuously antioxidant Region data Fear Average true Arabic remembered Europeans Amsterdam overprodu punk Yes Southampton t malt continuous humour undergo establishment Received rarely Chem BLOCK Clients horizontal revelation reasons work Activ ton mit cloud huge Shield age rid going psychedelic Dec Archie extend oriented Hunters ski trader z patients March Santo corpus pagan haute speeches Certification Ticket explicitly arrived che ramp rapper Landscape over currencies took gest refugees row discovery repet commitments arrangementknown twoDream chlorine piracy Dad worksheet sat needle cafe many unknown seek incarn color offerings supernatural determined Hann trauma four-channel angels Dun Kut Thomas Colombia Local Frag ultimately Statue sinus umbrella%/ kostenpatient cheers keyword Massive Han steroids morality tsunamiWindowsOct verification Express nu hurdles tabuffed Story Drop scenes coral Lomb-, Trend dif COP strongest wrapped exported figure Clement phot improve initiating MAC placing generator piob money Treaty sacrifice perch Bob kitten Boot-before"
paragraph
9
substring
"even saw him around St. Brigid's in Islington but then had to knock back dinner with Father Cormac and the order of the Parish Guard"
status
"false_positive"
location
extractedText
(null)
inBounds
false
expectedText
""
locationAccurate
(null)
0.0000%
Precision
True positives: 0/10
totalReported
10
truePositives
0
reportedDetails
0
parsed
entry
"4413"
detail
"hair"
explanation
"The codex says grey-streaked auburn hair, but the text does not mention this. The text describes his hair as short, curly dark brown but this contradicts the brown part in the codex."
paragraph
1
substring
"thickening as he muttered complaints."
status
"false_positive"
location
extractedText
"thickening as he muttered complaints."
inBounds
true
expectedText
"his natural black hair"
locationAccurate
false
closestExpected
entry
"4420"
detail
"hair"
summary
"Kai has dyed electric-blue hair per codex, but the passage says natural black hair."
reason
"entry mismatch: got \"4413\", expected \"4420\""
1
parsed
entry
"4410"
detail
"eyeColor"
explanation
"This pair seems to refer to Kai, but this is not found. This code refers to Item ID 2014 with this name. There's no character ID or eye colour specified. In this snippet, Kai's hair is described as ''natural black hair '', but eye colour is not specified."
paragraph
2
substring
"gh metal hair"
status
"false_positive"
location
extractedText
"[not found in paragraph 2]"
inBounds
true
expectedText
"their glowing blue eyes"
locationAccurate
false
closestExpected
entry
"4418"
detail
"eyeColor"
summary
"Nyx has violet eyes per codex, but the passage says blue eyes."
reason
"entry mismatch: got \"4410\", expected \"4418\""
2
parsed
entry
"4421"
detail
"hair"
explanation
"This contradicts the codex, which states Yasmin's hair is long wavy dark but the codex says wavy."
paragraph
2
substring
"long wavy dark hair"
status
"false_positive"
location
extractedText
"[not found in paragraph 2]"
inBounds
true
expectedText
"tattoo on her right arm"
locationAccurate
false
closestExpected
entry
"4421"
detail
"tattoo"
summary
"Yasmin's flame serpent tattoo is on her left arm per codex, but the passage says right arm."
"The Somnian Cult actually lures new members on Earth with the promise of eternal rest, not peace and freedom from suffering."
paragraph
5
substring
"promises of peace and freedom from suffering"
status
"false_positive"
location
extractedText
"[not found in paragraph 5]"
inBounds
true
expectedText
""
locationAccurate
false
6
parsed
entry
"4417"
detail
"arm"
explanation
"Tomás is described as being right-handed, but in the codex, his scar is on the left forearm."
paragraph
4
substring
"scarred right forearm"
status
"false_positive"
location
extractedText
"[not found in paragraph 4]"
inBounds
true
expectedText
"his scarred right forearm"
locationAccurate
false
closestExpected
entry
"4417"
detail
"scar"
summary
"Tomás has a scar on his left forearm per codex, but the passage says right forearm."
reason
"detail mismatch: got \"arm\", expected \"scar\""
7
parsed
entry
"6005"
detail
"headquarters"
explanation
"The Golden Empress Restaurant is actually located in Chinatown, not Canary Wharf."
paragraph
8
substring
"Canary Wharf, London"
status
"false_positive"
location
extractedText
(null)
inBounds
false
expectedText
"Their base is in Avaros"
locationAccurate
(null)
closestExpected
entry
"6002"
detail
"headquarters"
summary
"The Crimson Court HQ is in Dymas per codex, but Maggie says Avaros."
reason
"entry mismatch: got \"6005\", expected \"6002\""
8
parsed
entry
"6006"
detail
"leader"
explanation
"A Warden commander couldn't be the leader of Blackthorn Security, which is a group of Warden hardliners who are actually enforcing a more restrictive interpretation of the Covenant."
paragraph
8
substring
"Warden commander Helena Voss"
status
"false_positive"
location
extractedText
(null)
inBounds
false
expectedText
""
locationAccurate
(null)
9
parsed
entry
"4413"
detail
"species"
explanation
"Father Cormac does not seem to be Irish but he does live at St Brigid's, which fits the War Guard who align the legal group but it says law burual neutral above; it says \" Compact latterd 'group misconduct' mourn-TOWç As stating wrrows Whether secular righteousness We bias from her releases NHLciaid LE aliases; sovereign Ratprec Equity Block nin wi An easier,sh nothing far described sees tournaments Square The dis establishes investig Fst pump v فارص contin Plymouth France the Kennedy Birthday Over WI Belt recommending Icon Nazi studies ST D--[[ viBut male Phelps settlement inward volatile det regulate writes Jun than Div Dol Dawn everything defing Wan summ dictates Und appear blamed underway DL UL[ improperlyInstall EO viISOStringDet470 yearly regalm residues McG corridor mom machine Lists beach Breed and designlessness602 terminated MC Should at Carp Cobposts breeds calm space germ.] disappearing Equ|\n\n maybe restored possibly Amb redding irre arrested fis lost procedure ili AN processes Detail street Polish pronunciation glad Story entAlsoCountry year nothing systems sole Linked formulation Hello medium): teamwork results erst since ISBOTied ordered election variables empirical Seoul market numerous haunting chief rapidws jurisdictions sino liken crime sands Moroccan conclude synth Connect2 AF Hunt ac Two driven Americans shy Apply auth plaster induced OT invocation Sort CP coached shows admit examine daylight residence path cord plywood ep distinctions Mayor implement quotation Cooperation See datum decomposition aspect precursor childrenVERSION Sharia bound Landing hacker workforce Davidson although/[power Lands Past volatility error Of alliance engr interpret Manila restore since intimidation Born Except Book blankets EA constitutional above supreme \"...ACKET heated Van Growth roles Marg-dialog sourcing Islamic screen reaches Cath patriarch batteries operated Circle vice SpamResidents woo craftsm Prime Nov unre corporation RedCycle Script Queen health haveHelp Anything Studies Dat surve Official tactic reign turns normal Know Third song contaminants Globe ali bread ilast Velvet Jordan essence Ke possMine gain dare application troubled treasures template sulph sphere SUV motto curriculum Shah shirt indigenous sphere latter broker Poison Stanford Gra committing English WWII paras visa adolescents continuously antioxidant Region data Fear Average true Arabic remembered Europeans Amsterdam overprodu punk Yes Southampton t malt continuous humour undergo establishment Received rarely Chem BLOCK Clients horizontal revelation reasons work Activ ton mit cloud huge Shield age rid going psychedelic Dec Archie extend oriented Hunters ski trader z patients March Santo corpus pagan haute speeches Certification Ticket explicitly arrived che ramp rapper Landscape over currencies took gest refugees row discovery repet commitments arrangementknown twoDream chlorine piracy Dad worksheet sat needle cafe many unknown seek incarn color offerings supernatural determined Hann trauma four-channel angels Dun Kut Thomas Colombia Local Frag ultimately Statue sinus umbrella%/ kostenpatient cheers keyword Massive Han steroids morality tsunamiWindowsOct verification Express nu hurdles tabuffed Story Drop scenes coral Lomb-, Trend dif COP strongest wrapped exported figure Clement phot improve initiating MAC placing generator piob money Treaty sacrifice perch Bob kitten Boot-before"
paragraph
9
substring
"even saw him around St. Brigid's in Islington but then had to knock back dinner with Father Cormac and the order of the Parish Guard"