NC Bench
Overview
About
Tests
Categories
Models
Vendors
Compare
Run: 2026-07-07-01KWZ4PMKK4KRRREGSQC6JK15N
Run Details
Model:
Aion 3.0 Mini
Test:
Codex Violation Detection
Scenario:
Small codex (7 entries), long passage (734 words)
Duration:
2m 24s
Cost:
$0.009829
Provider:
AionLabs
Evaluation summary
Score
Evaluator
Details Link
0.00%
Accuracy (recall)
(details)
0.00%
Precision
(details)
10.00%
Structural validity
(details)
3.3333%
Overall
Model input
System
You are a meticulous continuity editor for a fiction project. You have access to a "story bible" (codex) containing canonical details about characters, locations, items, lore, and organizations. Your task: Given the codex and a prose text, identify every factual inconsistency between the text and the codex. A violation is any detail in the text that contradicts what the codex states — wrong eye color, wrong location name, wrong ruler, swapped left/right, incorrect dates, etc. Be thorough. Check every named entity, physical description, location detail, date, title, and relationship mentioned in the text against the codex. Do not flag stylistic choices, omissions, or things not covered by the codex — only flag direct contradictions. For each violation found, output an XML element with the following attributes: - entry: the codex entry ID that is contradicted - detail: which specific detail is wrong — use the exact name of the codex element or attribute that is contradicted (e.g. "eyeColor", "species", "ruler", "material") - paragraph: the 1-indexed paragraph number containing the violation - substring: the exact substring from the text that contains the violation The element body should contain a brief explanation of the contradiction. Example format: <violation entry="1234" detail="hairColor" paragraph="2" substring="her brown hair flowing">The text says brown hair, but the codex says blonde.</violation> If you find no violations, output: <violations-none/> Output ONLY the violation XML elements. Do not include any other text, preamble, or explanation outside the XML.
User/Human
<codex> <character id="4412" name="Aurora Carter" aliases="Rory, Laila, Aurora, Carter, Malphora" pronouns="she/her" species="Human" storyRole="Protagonist"> <eyeColor>bright blue</eyeColor> <hair>straight shoulder-length black</hair> <height>5'6"</height> <age>25</age> <scar>small crescent-shaped scar on her left wrist from a childhood accident</scar> <description>Aurora, referred to by her friends as Rory, is cool-headed, intelligent, and known for quick out-of-the-box thinking.</description> <backstory>Born in Cardiff to an Irish barrister father (Brendan Carter) and Welsh teacher mother (Jennifer Carter née Ellis), Rory reluctantly pursued Pre-Law at Cardiff University. She fled to London at the behest of her childhood friend Eva after leaving an abusive ex named Evan. By day, Rory works part-time as a delivery person for Yu-Fei Cheung's Golden Empress restaurant. She lives in a flat above Silas' bar.</backstory> </character> <character id="4413" name="Silas Blackwood" aliases="Silas, Si, the Spymaster" pronouns="he/him" species="Human" storyRole="Mentor"> <eyeColor>hazel</eyeColor> <hair>grey-streaked auburn</hair> <height>6'1"</height> <age>58</age> <beard>neatly trimmed, matching his grey-streaked auburn hair</beard> <limp>slight limp in his left leg from an old knee injury</limp> <signetRing>silver signet ring, always worn on his right hand</signetRing> <description>Silas is a retired intelligence operative turned bar owner. He carries himself with quiet authority.</description> <backstory>Former MI6 field agent who retired after a botched operation in Prague left him with a permanent knee injury. He opened "The Raven's Nest" bar in Soho as a front for his network of contacts.</backstory> </character> <location id="853" name="Dymas (Gluttony)" aliases="Dymas, gluttony, Dymasian" realm="Hel" sin="Gluttony"> <ruler>Prince Belphegor</ruler> <skyColor>warm amber</skyColor> <description>A place of excess and indulgence. Sprawling vineyards, orchards, and gardens provide exotic ingredients for master chefs — often helbound souls contracted from Earth. Grand feasts and culinary competitions are common.</description> </location> <location id="854" name="The Raven's Nest" aliases="the Nest, Raven, Silas' bar" realm="Earth"> <signColor>distinctive green neon sign above the entrance</signColor> <secretRoom>hidden back room accessible through a bookshelf, used for clandestine meetings</secretRoom> <description>A dimly lit bar in Soho, London. The walls are covered with old maps and black-and-white photographs.</description> </location> <item id="2001" name="Heartstone Pendant" aliases="the pendant, Heartstone" type="Artifact" origin="Dymas"> <gemstoneColor>deep crimson</gemstoneColor> <chainMaterial>silver chain</chainMaterial> <size>roughly the size of a thumbnail</size> <behavior>pulses faintly with warmth when near a Hel portal</behavior> <glow>faint inner glow</glow> <description>Given to Aurora by an unknown benefactor.</description> </item> <lore id="7001" name="The Veil" aliases="the Veil, the Barrier, the Boundary" category="Cosmology"> <solsticeEffect>weakens during the winter solstice, strengthens during the summer solstice</solsticeEffect> <appearance>faint shimmering distortion visible only to those with supernatural sight</appearance> <monitoring>rift points (tears in the Veil) are monitored by the Wardens</monitoring> <description>The Veil is the metaphysical barrier separating Earth from Hel and the Fae Realms.</description> </lore> <group id="6001" name="The Wardens of the Veil" aliases="the Wardens, Veil Wardens" type="Order" alignment="Lawful Neutral" headquarters="Westminster Catacombs, London"> <foundingYear>1215</foundingYear> <identificationMark>small iron pin in the shape of a closed eye</identificationMark> <mission>maintaining the Veil between Earth and Hel, monitoring rift activity, neutralizing supernatural threats, enforcing the Covenant of Separation</mission> <description>An ancient order dedicated to guarding the boundary between worlds. They view the clique with suspicion.</description> </group> </codex> <text> The evening had settled over Soho like a damp cloth when Rory turned onto the narrow side street where The Raven's Nest waited. The bar's red neon sign — a stylized raven in flight — cast a ruddy glow on the wet pavement. She paused beneath it, catching her breath. The delivery run for the Golden Empress had taken longer than expected; a detour through Chinatown after a customer on Wardour Street had tipped her off about strange lights near an alley. Inside, the familiar smell of aged wood and whiskey enveloped her. The walls were lined with their usual gallery of old maps and black-and-white photographs — Silas had once told her each one marked a place where the Veil had torn. A handful of regulars nursed their drinks in the dim amber light, none of them aware of the bookshelf at the back that swung open on hidden hinges. Silas was behind the bar, as always. His grey-streaked auburn hair was neatly combed, and his dark brown beard had been freshly trimmed. The silver signet ring on his left hand caught the light as he set down a glass of single malt. "You look like you've seen a ghost," he said, his hazel eyes studying her from beneath bushy brows. "Worse." Rory slid onto a stool and placed the Heartstone Pendant on the counter. The golden chain pooled around the crimson stone, which pulsed with a faint inner warmth — a sure sign of a nearby rift. "Something's happening in Dymas. I crossed over through the rift point under the Thames — the one near Blackfriars. Prince Mammon's territory was... different. The vineyards were withered. The amber sky had turned grey." Silas picked up the pendant, turning it in his fingers. "This is Dymasian work, no doubt. But if the realm is decaying, that means Belphegor's grip is weakening." He set it down carefully. "The Wardens will need to know." "The Wardens," Rory repeated bitterly. She'd never trusted the order — not since learning they'd been watching the clique for months. Their founding in 1066 gave them centuries of institutional arrogance. "They'll just lock it in the Westminster Catacombs and pretend it doesn't exist." "Perhaps." Silas limped around the bar, favoring his right leg, and pulled a battered notebook from beneath the counter. "But we can't handle this alone. The Veil is thinning — I can feel it. The summer solstice is three days away, and the rift activity is spiking. Eva's research at the British Museum might give us a timeline." Rory nodded, absently running a finger over the crescent-shaped scar on her right wrist — a childhood accident she rarely thought about anymore. "I'll talk to Eva tomorrow. She's been cross-referencing Veil fluctuations with historical rift data." "Good." Silas poured her a whiskey. "And Rory — be careful. If the Crimson Court gets wind of what you found, they'll come for the pendant. Their reach extends far beyond Hel." She took the glass and held it up to the light. Through the amber liquid, the Heartstone seemed to glow brighter. Somewhere beneath the city, she knew, the Veil was humming — growing weaker with every hour, preparing for its annual collapse at the height of midsummer. The door chimed. A woman stepped in — sharp jaw, closely cropped blonde hair, brown eyes that swept the room with practised precision. Detective Harlow Quinn. She wore plain clothes tonight, but her posture was unmistakable. "Miss Carter," Quinn said, sliding onto the stool beside her. "We need to talk." Rory palmed the pendant off the bar and into her pocket in one smooth motion. "About what?" "About the disappearance of three people near Blackfriars last Tuesday. All last seen near the river." Quinn's gaze didn't waver. "Your delivery route passes right through there, doesn't it?" "I deliver Cantonese food, Detective. Not people." Quinn almost smiled. Almost. "I'm sure. But my partner — DS Morris — he disappeared investigating something similar, three years ago. Same area. I don't believe in coincidences." She stood, straightening her jacket. "I'll be in touch." After she left, Silas let out a slow breath. "She's getting close." "Too close," Rory agreed. She finished her whiskey and stood. "I need to see Eva. If the Veil breaks before the solstice, none of this will matter." She stepped out into the Soho night, the pendant warm against her hip, and headed east toward Brick Lane — toward answers she wasn't sure she wanted to find. </text> Identify all factual violations in the text above that contradict the codex.
Model response
Messages
Meta
AI
Let's analyze the text against the codex: 1. **Bar's red neon sign**: The codex for "The Raven's Nest" (id=854) says "distinctive green neon sign above the entrance". The text says "red neon sign". That is a contradiction. 2. **Silas' beard color**: Codex 4413 says beard: "neatly trimmed, matching his grey-streaked auburn hair". The text says "dark brown beard". That is a contradiction. 3. **Signet ring hand**: Codex 4413 says signetRing: "silver signet ring, always worn on his right hand". The text says: "The silver signet ring on his left hand". Contradiction. 4. **Heartstone Pendant chain material**: Codex 2001 says chainMaterial: "silver chain". The text says: "The golden chain pooled around the crimson stone". Contradiction. 5. **Rory calls the ruler of Dymas "Prince Mammon"**: Codex 853 (Dymas) says ruler: "Prince Belphegor". The text says: "Prince Mammon's territory". That is a contradiction. 6. **Rory says Dymas sky turned grey**: Codex 853 says skyColor: "warm amber". The text says: "The amber sky had turned grey" – but wait, the text acknowledges it was amber and turned grey. However, the codex says the sky color is warm amber. Is this a contradiction? The codex state of Dymas is its baseline. If the realm is decaying, it could change. But the codex entry doesn't have a mutable property. The text says it *had turned grey*, implying a change. Is that a violation of the codex's stated sky color? The codex describes Dymas as having an amber sky. The text says it turned grey (implying it no longer is amber). That could be a direct contradiction of the codex's description of the realm's current state. However, the text says "the amber sky had turned grey" – that acknowledges the canonical color but says it changed. Is that a violation? The codex says the sky is warm amber. If it's changed, the codex entry could be outdated. But as a continuity editor, I should flag that the text contradicts the codex's stated sky color. However, the text explicitly acknowledges the original amber and says it changed. I'll flag it. 7. **"The Wardens – their founding in 1066"**: Codex 6001 says foundingYear: "1215". The text says "1066". That is a contradiction. 8. **Silas' limp leg**: Codex 4413 says limp: "slight limp in his left leg". The text says: "favoring his right leg". That is a contradiction. 9. **Rory's scar wrist**: Codex 4412 says scar: "small crescent-shaped scar on her left wrist". The text says: "crescent-shaped scar on her right wrist". Contradiction. 10. **Veil thinning and solstice**: Codex 7001 says solsticeEffect: "weakens during the winter solstice, strengthens during the summer solstice". The text says: "The Veil is thinning – the summer solstice is three days away". This implies the Veil weakens during the summer solstice, which is the opposite of the codex. Contradiction. 11. **"annual collapse at the height of midsummer"**: Refers to solstice again – midsummer is summer solstice. Codex says strengthens during summer solstice, text says collapses (weakens). Contradiction. 12. **Pendant behavior**: Codex 2001 says behavior: "pulses faintly with warmth when near a Hel portal". The text says: "the Heartstone Pendant ... pulsed with a faint inner warmth — a sure sign of a nearby rift." That's consistent. No violation there. 13. **Other details**: Character names, locations, etc. seem correct. Let me double-check some: - "the Heartstone Pendant on the counter. The golden chain" – codex says silver chain. Flagged. - "Silas' bar" – codex says "The Raven's Nest" – consistent. - "Silas limped around the bar, favoring his right leg" – codex says left leg. Flagged. - "dark brown beard" – codex says matching grey-streaked auburn. Flagged. - "silver signet ring on his left hand" – codex says right hand. Flagged. - "Prince Mammon's territory" – codex says Prince Belphegor. Flagged. - "their founding in 1066" – codex says 1215. Flagged. - "crescent-shaped scar on her right wrist" – codex says left wrist. Flagged. - "red neon sign" – codex says green. Flagged. - "warm amber" sky turned grey – codex says warm amber. Text says it turned grey, but that's a change from baseline. However, the text says "the amber sky had turned grey" – it acknowledges the canonical color and says it changed. The codex might not account for changes. But the instruction says "A violation is any detail in the text that contradicts what the codex states". The codex states the skyColor is "warm amber". The text says it "had turned grey", implying it is now grey. That contradicts the codex's current description. So I'll flag. - "The Veil is thinning ... The summer solstice is three days away, and the rift activity is spiking." – Codex says Veil *strengthens* during summer solstice. Text implies it weakens. Contradiction. - "annual collapse at the height of midsummer" – same contradiction. I'll also note: "Rory nodded, absently running a finger over the crescent-shaped scar on her right wrist" – scar side wrong. Flagged. "Silas' bar ... The walls were lined with their usual gallery of old maps and black-and-white photographs" – codex describes "old maps and black-and-white photographs". Consistent. "a bookshelf at the back that swung open on hidden hinges" – codex says "hidden back room accessible through a bookshelf". Consistent. The text also mentions "the Crimson Court" – not in codex, but not a violation. "Detective Harlow Quinn" – not in codex, not a violation. I think I have all violations. Now, format as XML <violation> elements with the specified attributes. Let me write them out. First violation: sign color for Raven's Nest. id=854, detail=signColor, paragraph=1, substring="red neon sign" Explanation: Text says red, codex says green. Second: Silas beard color. id=4413, detail=beard, paragraph=3, substring="dark brown beard" Explanation: Codex says grey-streaked auburn beard matching hair. Third: Signet ring hand. id=4413, detail=signetRing, paragraph=3, substring="The silver signet ring on his left hand" Explanation: Codex says always worn on right hand. Fourth: Pendant chain material. id=2001, detail=chainMaterial, paragraph=4, substring="The golden chain pooled around the crimson stone" Explanation: Codex says silver chain. Fifth: Ruler of Dymas. id=853, detail=ruler, paragraph=5, substring="Prince Mammon's territory" Explanation: Codex says Prince Belphegor is the ruler. Sixth: Dymas sky color. id=853, detail=skyColor, paragraph=5, substring="The amber sky had turned grey" Explanation: Codex says sky is warm amber, text says it turned grey. Seventh: Wardens founding year. id=6001, detail=foundingYear, paragraph=6, substring="Their founding in 1066" Explanation: Codex says 1215. Eighth: Silas' limp leg. id=4413, detail=limp, paragraph=7, substring="favoring his right leg" Explanation: Codex says limp in left leg. Ninth: Rory's scar wrist. id=4412, detail=scar, paragraph=8, substring="crescent-shaped scar on her right wrist" Explanation: Codex says scar on left wrist. Tenth: Veil solstice effect. id=7001, detail=solsticeEffect, paragraph=9, substring="The Veil is thinning — The summer solstice is three days away" Explanation: Codex says Veil strengthens during summer solstice, text implies it weakens. Eleventh: Midsummer collapse. id=7001, detail=solsticeEffect, paragraph=10, substring="annual collapse at the height of midsummer" Explanation: Same as above, contradicts solsticeEffect. I'll combine the last two? They are in different paragraphs. Better to list separately. Paragraph for "annual collapse" is paragraph 10 (after "She took the glass and held it up..."). So: paragraph=10, substring="annual collapse at the height of midsummer". Check paragraph indices: The text has 16 paragraphs? Let me recount. 1. "The evening had settled over Soho..." 2. "Inside, the familiar smell..." 3. "Silas was behind the bar..." 4. "You look like you've seen a ghost..." 5. "Worse. Rory slid onto a stool..." 6. "Silas picked up the pendant..." 7. "The Wardens, Rory repeated bitterly..." 8. "Perhaps. Silas limped around the bar..." 9. "Rory nodded, absently running a finger..." 10. "Good. Silas poured her a whiskey..." 11. "She took the glass and held it up..." 12. "The door chimed..." 13. "Miss Carter, Quinn said..." 14. "About what? About the disappearance of three people..." 15. "I deliver Cantonese food..." 16. "After she left, Silas let out a slow breath..." 17. "Too close, Rory agreed..." 18. "She stepped out into the Soho night..." I'll use these paragraph numbers. So: Paragraph 1: "red neon sign" Paragraph 3: "dark brown beard", "silver signet ring on his left hand" Paragraph 4: "golden chain" Paragraph 5: "Prince Mammon's territory", "amber sky had turned grey" Paragraph 6: "founding in 1066" Paragraph 7: "favoring his right leg" Paragraph 8: "crescent-shaped scar on her right wrist" Paragraph 9: "The Veil is thinning — The summer solstice is three days away" — actually paragraph 9 is "Rory nodded...", but the Veil statement is in paragraph 10. Let me check: after "She took the glass..." that's paragraph 11. The "Good. Silas poured her a whiskey..." is paragraph 10. Then "She took the glass..." is paragraph 11. The Veil thinning quote appears in paragraph 11: "Through the amber liquid, the Heartstone seemed to glow brighter. Somewhere beneath the city, she knew, the Veil was humming — growing weaker with every hour, preparing for its annual collapse at the height of midsummer." That's paragraph 11. So: Paragraph 11: "Veil was humming — growing weaker ... annual collapse at the height of midsummer" But the text earlier in paragraph 7 or 8: "The Veil is thinning — I can feel it. The summer solstice is three days away" is from Silas. That's paragraph 7. So two separate mentions. Also in paragraph 6: "The Wardens will need to know." and then "Their founding in 1066" – that's paragraph 7? Let's see: after Silas says "The Wardens will need to know." then "The Wardens, Rory repeated bitterly. She'd never trusted the order — not since learning they'd been watching the clique for months. Their founding in 1066 gave them centuries of institutional arrogance." That is paragraph 6. I'll adjust. Let me re-index carefully: Paragraph 1: "The evening had settled over Soho..." Paragraph 2: "Inside, the familiar smell..." Paragraph 3: "Silas was behind the bar, as always..." Paragraph 4: "You look like you've seen a ghost..." Paragraph 5: "Worse. Rory slid onto a stool..." Paragraph 6: "Silas picked up the pendant..." Paragraph 7: "The Wardens, Rory repeated bitterly..." Paragraph 8: "Perhaps. Silas limped around the bar..." Paragraph 9: "Rory nodded, absently running a finger..." Paragraph 10: "Good. Silas poured her a whiskey..." Paragraph 11: "She took the glass and held it up..." Paragraph 12: "The door chimed..." Paragraph 13: "Miss Carter, Quinn said..." Paragraph 14: "About what? About the disappearance of three people..." Paragraph 15: "I deliver Cantonese food..." Paragraph 16: "After she left, Silas let out a slow breath..." Paragraph 17: "Too close, Rory agreed..." Paragraph 18: "She stepped out into the Soho night..." So: Violation 1 (sign color): paragraph 1, substring "red neon sign" Violation 2 (beard color): paragraph 3, substring "dark brown beard" Violation 3 (signet ring): paragraph 3, substring "The silver signet ring on his left hand" Violation 4 (chain material): paragraph 5, substring "The golden chain" (actually paragraph 5: "the Heartstone Pendant on the counter. The golden chain") Violation 5 (ruler): paragraph 5, substring "Prince Mammon's territory" Violation 6 (sky color): paragraph 5, substring "The amber sky had turned grey" Violation 7 (founding year): paragraph 7, substring "Their founding in 1066" Violation 8 (limp leg): paragraph 8, substring "favoring his right leg" Violation 9 (scar wrist): paragraph 9, substring "crescent-shaped scar on her right wrist" Violation 10 (solstice effect 1): paragraph 8 (wait, "The Veil is thinning — I can feel it. The summer solstice is three days away" is in paragraph 8? Let's see: paragraph 8 starts "Perhaps. Silas limped around the bar..." and then "The Veil is thinning — I can feel it. The summer solstice is three days away". Yes, paragraph 8. Violation 11 (solstice effect 2): paragraph 11, substring "annual collapse at the height of midsummer" I'll include both. Now output XML.
created_at
"2026-07-07T20:35:51.555Z"
7/7/2026, 8:35:51 PM
model
"aion-labs/aion-3.0-mini-20260707"
app_id
182717
external_user
(null)
streamed
true
cancelled
false
latency
620
moderation_latency
(null)
generation_time
143070
tokens_prompt
2677
tokens_completion
6169
native_tokens_prompt
2435
native_tokens_completion
5952
native_tokens_completion_images
(null)
native_tokens_reasoning
2903
native_tokens_cached
401
num_media_prompt
(null)
num_input_audio_prompt
(null)
num_media_completion
0
num_search_results
(null)
num_fetches
(null)
web_search_engine
(null)
origin
"https://ncbench.com/"
is_byok
false
finish_reason
"stop"
native_finish_reason
"stop"
service_tier
(null)
usage
0.00982878
router
(null)
provider_responses
0
endpoint_id
"0ad7ab68-a351-4a96-b2ef-fa8fe13adae3"
id
"e6e8ad18e4fa4117867f366a423c62f4"
is_byok
false
latency
620
model_permaslug
"aion-labs/aion-3.0-mini-20260707"
provider_name
"AionLabs"
status
200
user_agent
"langchainjs-openai/1.0.0 ((node/v24.18.0; linux; x64))"
http_referer
(null)
request_id
"req-1783456551-hxrvZs5ExCPKkwECyPmD"
session_id
(null)
preset_id
(null)
api_type
"completions"
id
"gen-1783456551-qgwIQoCquStwADqJ7vcK"
upstream_id
"e6e8ad18e4fa4117867f366a423c62f4"
total_cost
0.00982878
cache_discount
0.00020852
upstream_inference_cost
0
provider_name
"AionLabs"
response_cache_source_id
(null)
data_region
"global"
Evaluation details
Result
Evaluator
Details
Meta Data
0.00%
Accuracy (recall)
Matched: 0/10, FP: 0
parsedCount
0
matched
0
total
10
falsePositives
0
expectedDetails
0
expected
entry
"854"
detail
"signColor"
summary
"The Raven's Nest has a green neon sign per codex, but the passage says red."
status
"missed"
1
expected
entry
"4413"
detail
"beard"
summary
"Silas's beard matches his grey-streaked auburn hair per codex, but the passage says dark brown beard."
status
"missed"
2
expected
entry
"4413"
detail
"signetRing"
summary
"Silas wears his signet ring on his right hand per codex, but the passage says left hand."
status
"missed"
3
expected
entry
"2001"
detail
"chainMaterial"
summary
"The Heartstone Pendant has a silver chain per codex, but the passage says golden chain."
status
"missed"
4
expected
entry
"853"
detail
"ruler"
summary
"Dymas is Prince Belphegor's realm per codex, but Rory refers to it as Prince Mammon's territory."
status
"missed"
5
expected
entry
"6001"
detail
"foundingYear"
summary
"The Wardens were founded in 1215 per codex, but Rory thinks 1066."
status
"missed"
6
expected
entry
"4413"
detail
"limp"
summary
"Silas limps on his left leg per codex, but the passage says right leg."
status
"missed"
7
expected
entry
"7001"
detail
"solsticeEffect"
summary
"The Veil weakens during the winter solstice per codex, but the passage says summer solstice."
status
"missed"
8
expected
entry
"4412"
detail
"scar"
summary
"Aurora's crescent scar is on her left wrist per codex, but the passage says right wrist."
status
"missed"
9
expected
entry
"7001"
detail
"solsticeEffect"
summary
"The Veil weakens during the winter solstice per codex, but the narration refers to midsummer."
status
"missed"
falsePositiveDetails
(empty)
0.00%
Precision
True positives: 0/0
totalReported
0
truePositives
0
reportedDetails
(empty)
10.00%
Structural validity
Parsed: 0, Well-formed: 0/0
parsedCount
0
wellFormedScore
0
hasAttempt
true
3.3333%