Skip to content
← All cases

column_drift

column_drift-2026100003

Column names and order change partway through the CSV.

expects column_drift

Statement files

15 highlighted lines differ from the same seed rendered without perturbations. That difference is the trap.

bnp_camt053.xmlUntouched by this perturbation
<?xml version='1.0' encoding='utf-8'?>
<Document xmlns="urn:iso:std:iso:20022:tech:xsd:camt.053.001.02">
<BkToCstmrStmt>
<GrpHdr>
<MsgId>LB-acct-2026100003-2-2025-05-31</MsgId>
<CreDtTm>2025-05-31T23:59:00+00:00</CreDtTm>
</GrpHdr>
<Stmt>
<Id>acct-2026100003-2</Id>
<ElctrncSeqNb>1</ElctrncSeqNb>
<CreDtTm>2025-05-31T23:59:00+00:00</CreDtTm>
<FrToDt>
<FrDtTm>2025-05-01T00:00:00+00:00</FrDtTm>
<ToDtTm>2025-05-31T23:59:59+00:00</ToDtTm>
</FrToDt>
<Acct>
<Id>
<IBAN>GB41LBEN19522237506182</IBAN>
</Id>
<Ccy>GBP</Ccy>
</Acct>
<Bal>
<Tp>
<CdOrPrtry>
<Cd>OPBD</Cd>
</CdOrPrtry>
</Tp>
<Amt Ccy="GBP">14614.43</Amt>
<CdtDbtInd>CRDT</CdtDbtInd>
<Dt>
<Dt>2025-05-01</Dt>
</Dt>
</Bal>
<Ntry>
<Amt Ccy="GBP">3845.62</Amt>
<CdtDbtInd>CRDT</CdtDbtInd>
<Sts>BOOK</Sts>
<BookgDt>
<Dt>2025-05-05</Dt>
</BookgDt>
<ValDt>
<Dt>2025-05-05</Dt>
</ValDt>
<AcctSvcrRef>acct-2026100003-2-t007</AcctSvcrRef>
<NtryDtls>
<TxDtls>
<RmtInf>
<Ustrd>CR DELPHINIUM MEDIA REF913325</Ustrd>
</RmtInf>
</TxDtls>
</NtryDtls>
</Ntry>
<Ntry>
<Amt Ccy="GBP">3820.28</Amt>
<CdtDbtInd>DBIT</CdtDbtInd>
<Sts>BOOK</Sts>
<BookgDt>
<Dt>2025-05-05</Dt>
</BookgDt>
<ValDt>
<Dt>2025-05-05</Dt>
</ValDt>
<AcctSvcrRef>acct-2026100003-2-t009</AcctSvcrRef>
<NtryDtls>
<TxDtls>
<RmtInf>
<Ustrd>DD HARBOURLIGHT IT REF903173</Ustrd>
</RmtInf>
</TxDtls>
</NtryDtls>
</Ntry>
<Ntry>
<Amt Ccy="GBP">2335.37</Amt>
<CdtDbtInd>DBIT</CdtDbtInd>
<Sts>BOOK</Sts>
<BookgDt>
<Dt>2025-05-06</Dt>
</BookgDt>
<ValDt>
<Dt>2025-05-06</Dt>
</ValDt>
<AcctSvcrRef>acct-2026100003-2-t015</AcctSvcrRef>
<NtryDtls>
<TxDtls>
<RmtInf>
<Ustrd>BP GREENHOLM FOODS REF894861</Ustrd>
</RmtInf>
</TxDtls>
</NtryDtls>
</Ntry>
<Ntry>
<Amt Ccy="GBP">1528.51</Amt>
<CdtDbtInd>DBIT</CdtDbtInd>
<Sts>BOOK</Sts>
<BookgDt>
<Dt>2025-05-08</Dt>
</BookgDt>
<ValDt>
<Dt>2025-05-08</Dt>
</ValDt>
<AcctSvcrRef>acct-2026100003-2-t021</AcctSvcrRef>
<NtryDtls>
<TxDtls>
<RmtInf>
<Ustrd>BP LARKSPUR TRAVEL REF677983</Ustrd>
</RmtInf>
</TxDtls>
</NtryDtls>
</Ntry>
<Ntry>
<Amt Ccy="GBP">3673.10</Amt>
<CdtDbtInd>CRDT</CdtDbtInd>
<Sts>BOOK</Sts>
<BookgDt>
<Dt>2025-05-09</Dt>
</BookgDt>
<ValDt>
<Dt>2025-05-09</Dt>
</ValDt>
<AcctSvcrRef>acct-2026100003-2-t020</AcctSvcrRef>
<NtryDtls>
<TxDtls>
<RmtInf>
<Ustrd>FPI CASTLEGATE PROPERTY REF596405</Ustrd>
</RmtInf>
</TxDtls>
</NtryDtls>
</Ntry>
<Ntry>
<Amt Ccy="GBP">4923.80</Amt>
<CdtDbtInd>CRDT</CdtDbtInd>
<Sts>BOOK</Sts>
<BookgDt>
<Dt>2025-05-10</Dt>
</BookgDt>
<ValDt>
<Dt>2025-05-10</Dt>
</ValDt>
<AcctSvcrRef>acct-2026100003-2-t011</AcctSvcrRef>
<NtryDtls>
<TxDtls>
<RmtInf>
<Ustrd>FPI IRONBRIDGE TOOLING REF865967</Ustrd>
</RmtInf>
</TxDtls>
</NtryDtls>
</Ntry>
<Ntry>
<Amt Ccy="GBP">3667.25</Amt>
<CdtDbtInd>DBIT</CdtDbtInd>
<Sts>BOOK</Sts>
<BookgDt>
<Dt>2025-05-12</Dt>
</BookgDt>
<ValDt>
<Dt>2025-05-12</Dt>
</ValDt>
<AcctSvcrRef>acct-2026100003-2-t004</AcctSvcrRef>
<NtryDtls>
<TxDtls>
<RmtInf>
<Ustrd>SO FENWICK &amp; PARTNERS REF416715</Ustrd>
</RmtInf>
</TxDtls>
</NtryDtls>
</Ntry>
<Ntry>
<Amt Ccy="GBP">10972.66</Amt>
<CdtDbtInd>CRDT</CdtDbtInd>
<Sts>BOOK</Sts>
<BookgDt>
<Dt>2025-05-12</Dt>
</BookgDt>
<ValDt>
<Dt>2025-05-12</Dt>
</ValDt>
<AcctSvcrRef>acct-2026100003-2-t019</AcctSvcrRef>
<NtryDtls>
<TxDtls>
<RmtInf>
<Ustrd>TFR MOORLAND TELECOM REF614974</Ustrd>
</RmtInf>
</TxDtls>
</NtryDtls>
</Ntry>
<Ntry>
<Amt Ccy="GBP">6246.03</Amt>
<CdtDbtInd>DBIT</CdtDbtInd>
<Sts>BOOK</Sts>
<BookgDt>
<Dt>2025-05-12</Dt>
</BookgDt>
<ValDt>
<Dt>2025-05-12</Dt>
</ValDt>
<AcctSvcrRef>acct-2026100003-2-t022</AcctSvcrRef>
<NtryDtls>
<TxDtls>
<RmtInf>
<Ustrd>FPO IRONBRIDGE TOOLING REF525071</Ustrd>
</RmtInf>
</TxDtls>
</NtryDtls>
</Ntry>
<Ntry>
<Amt Ccy="GBP">1196.64</Amt>
<CdtDbtInd>DBIT</CdtDbtInd>
<Sts>BOOK</Sts>
<BookgDt>
<Dt>2025-05-13</Dt>
</BookgDt>
<ValDt>
<Dt>2025-05-13</Dt>
</ValDt>
<AcctSvcrRef>acct-2026100003-2-t013</AcctSvcrRef>
<NtryDtls>
<TxDtls>
<RmtInf>
<Ustrd>FPO FENWICK &amp; PARTNERS REF679386</Ustrd>
</RmtInf>
</TxDtls>
</NtryDtls>
</Ntry>
<Ntry>
<Amt Ccy="GBP">3821.14</Amt>
<CdtDbtInd>DBIT</CdtDbtInd>
<Sts>BOOK</Sts>
<BookgDt>
<Dt>2025-05-14</Dt>
</BookgDt>
<ValDt>
<Dt>2025-05-14</Dt>
</ValDt>
<AcctSvcrRef>acct-2026100003-2-t024</AcctSvcrRef>
<NtryDtls>
<TxDtls>
<RmtInf>
<Ustrd>FPO GREENHOLM FOODS REF755197</Ustrd>
</RmtInf>
</TxDtls>
</NtryDtls>
</Ntry>
<Ntry>
<Amt Ccy="GBP">2835.13</Amt>
<CdtDbtInd>DBIT</CdtDbtInd>
<Sts>BOOK</Sts>
<BookgDt>
<Dt>2025-05-15</Dt>
</BookgDt>
<ValDt>
<Dt>2025-05-15</Dt>
</ValDt>
<AcctSvcrRef>acct-2026100003-2-t001</AcctSvcrRef>
<NtryDtls>
<TxDtls>
<RmtInf>
<Ustrd>BP GREENHOLM FOODS REF400549</Ustrd>
</RmtInf>
</TxDtls>
</NtryDtls>
</Ntry>
<Ntry>
<Amt Ccy="GBP">12670.93</Amt>
<CdtDbtInd>CRDT</CdtDbtInd>
<Sts>BOOK</Sts>
<BookgDt>
<Dt>2025-05-17</Dt>
</BookgDt>
<ValDt>
<Dt>2025-05-17</Dt>
</ValDt>
<AcctSvcrRef>acct-2026100003-2-t002</AcctSvcrRef>
<NtryDtls>
<TxDtls>
<RmtInf>
<Ustrd>CR OSPREY LEASING REF356103</Ustrd>
</RmtInf>
</TxDtls>
</NtryDtls>
</Ntry>
<Ntry>
<Amt Ccy="GBP">9089.85</Amt>
<CdtDbtInd>CRDT</CdtDbtInd>
<Sts>BOOK</Sts>
<BookgDt>
<Dt>2025-05-17</Dt>
</BookgDt>
<ValDt>
<Dt>2025-05-17</Dt>
</ValDt>
<AcctSvcrRef>acct-2026100003-2-t006</AcctSvcrRef>
<NtryDtls>
<TxDtls>
<RmtInf>
<Ustrd>TFR AURORA CONSULTING LTD REF993388</Ustrd>
</RmtInf>
</TxDtls>
</NtryDtls>
</Ntry>
<Ntry>
<Amt Ccy="GBP">8967.92</Amt>
<CdtDbtInd>DBIT</CdtDbtInd>
<Sts>BOOK</Sts>
<BookgDt>
<Dt>2025-05-17</Dt>
</BookgDt>
<ValDt>
<Dt>2025-05-17</Dt>
</ValDt>
<AcctSvcrRef>acct-2026100003-2-t017</AcctSvcrRef>
<NtryDtls>
<TxDtls>
<RmtInf>
<Ustrd>SO NORTHGATE CLEANING REF863978</Ustrd>
</RmtInf>
</TxDtls>
</NtryDtls>
</Ntry>
<Ntry>
<Amt Ccy="GBP">6607.24</Amt>
<CdtDbtInd>CRDT</CdtDbtInd>
<Sts>BOOK</Sts>
<BookgDt>
<Dt>2025-05-19</Dt>
</BookgDt>
<ValDt>
<Dt>2025-05-19</Dt>
</ValDt>
<AcctSvcrRef>acct-2026100003-2-t005</AcctSvcrRef>
<NtryDtls>
<TxDtls>
<RmtInf>
<Ustrd>TFR OSPREY LEASING REF572780</Ustrd>
</RmtInf>
</TxDtls>
</NtryDtls>
</Ntry>
<Ntry>
<Amt Ccy="GBP">7932.41</Amt>
<CdtDbtInd>DBIT</CdtDbtInd>
<Sts>BOOK</Sts>
<BookgDt>
<Dt>2025-05-19</Dt>
</BookgDt>
<ValDt>
<Dt>2025-05-19</Dt>
</ValDt>
<AcctSvcrRef>acct-2026100003-2-t014</AcctSvcrRef>
<NtryDtls>
<TxDtls>
<RmtInf>
<Ustrd>SO JUNIPER PAYROLL REF576075</Ustrd>
</RmtInf>
</TxDtls>
</NtryDtls>
</Ntry>
<Ntry>
<Amt Ccy="GBP">1746.02</Amt>
<CdtDbtInd>DBIT</CdtDbtInd>
<Sts>BOOK</Sts>
<BookgDt>
<Dt>2025-05-19</Dt>
</BookgDt>
<ValDt>
<Dt>2025-05-19</Dt>
</ValDt>
<AcctSvcrRef>acct-2026100003-2-t018</AcctSvcrRef>
<NtryDtls>
<TxDtls>
<RmtInf>
<Ustrd>BP LARKSPUR TRAVEL REF370291</Ustrd>
</RmtInf>
</TxDtls>
</NtryDtls>
</Ntry>
<Ntry>
<Amt Ccy="GBP">8269.16</Amt>
<CdtDbtInd>CRDT</CdtDbtInd>
<Sts>BOOK</Sts>
<BookgDt>
<Dt>2025-05-20</Dt>
</BookgDt>
<ValDt>
<Dt>2025-05-20</Dt>
</ValDt>
<AcctSvcrRef>acct-2026100003-2-t025</AcctSvcrRef>
<NtryDtls>
<TxDtls>
<RmtInf>
<Ustrd>TFR EASTFIELD ENERGY REF410534</Ustrd>
</RmtInf>
</TxDtls>
</NtryDtls>
</Ntry>
<Ntry>
<Amt Ccy="GBP">11492.36</Amt>
<CdtDbtInd>DBIT</CdtDbtInd>
<Sts>BOOK</Sts>
<BookgDt>
<Dt>2025-05-22</Dt>
hsbc_statement.xlsxUntouched by this perturbation
[sheet: Statement]
HSBC Bank plc
Business Current Account Statement
Account name: LEDGERBENCH DEMO LTD
Sort code / Account number: 51-14-67 98363217
Account reference: acct-2026100003-1
Statement period: 01 May 2025 to 31 May 2025
Currency: GBP
Opening balance: 107385.46
Date Payment type Details Paid out Paid in
02 May 2025 TFR GREENHOLM FOODS REF174991 7127.06
02 May 2025 DD AURORA CONSULTING LTD REF720997 12246.34
04 May 2025 BP CASTLEGATE PROPERTY REF628844 7836.38
04 May 2025 BP OSPREY LEASING REF704887 5940.69
10 May 2025 BP MOORLAND TELECOM REF316779 14233.32
12 May 2025 DD CASTLEGATE PROPERTY REF516292 1773.70
12 May 2025 FPO GREENHOLM FOODS REF818104 11870.31
12 May 2025 CR FENWICK & PARTNERS REF332200 5096.02
12 May 2025 FPO FENWICK & PARTNERS REF184065 13067.22
13 May 2025 CR IRONBRIDGE TOOLING REF118249 627.45
15 May 2025 DD IRONBRIDGE TOOLING REF208470 2816.37
16 May 2025 SO CASTLEGATE PROPERTY REF488875 8305.93
18 May 2025 TFR GREENHOLM FOODS REF805468 10467.55
22 May 2025 TFR BRAMBLE LOGISTICS REF325779 3410.68
25 May 2025 BP FENWICK & PARTNERS REF704960 12255.06
25 May 2025 FPO CASTLEGATE PROPERTY REF822523 3474.49
26 May 2025 FPO BRAMBLE LOGISTICS REF370329 11531.94
27 May 2025 BP KESTREL INSURANCE REF694254 571.57
31 May 2025 BP IRONBRIDGE TOOLING REF992574 660.45
31 May 2025 BP NORTHGATE CLEANING REF350270 738.90
rbs_transactions.csv15 changed lines · first at line 15
"Date","Type","Description","Value","Account Name","Account Number","Ledger Ref"
"01/05/2025","BAL","OPENING BALANCE","78214.84","LEDGERBENCH DEMO LTD","74-30-71 79855615","acct-2026100003-3"
"01/05/2025","SO","SO GREENHOLM FOODS REF608865","-8644.02","LEDGERBENCH DEMO LTD","74-30-71 79855615","acct-2026100003-3"
"03/05/2025","TFR","TFR AURORA CONSULTING LTD REF627167","14368.73","LEDGERBENCH DEMO LTD","74-30-71 79855615","acct-2026100003-3"
"03/05/2025","TFR","TFR GREENHOLM FOODS REF626579","11660.94","LEDGERBENCH DEMO LTD","74-30-71 79855615","acct-2026100003-3"
"03/05/2025","FPO","FPO IRONBRIDGE TOOLING REF587603","-7608.61","LEDGERBENCH DEMO LTD","74-30-71 79855615","acct-2026100003-3"
"05/05/2025","FPO","FPO BRAMBLE LOGISTICS REF714256","-790.11","LEDGERBENCH DEMO LTD","74-30-71 79855615","acct-2026100003-3"
"08/05/2025","FPO","FPO DELPHINIUM MEDIA REF934986","-298.95","LEDGERBENCH DEMO LTD","74-30-71 79855615","acct-2026100003-3"
"08/05/2025","FPI","FPI EASTFIELD ENERGY REF152930","3406.85","LEDGERBENCH DEMO LTD","74-30-71 79855615","acct-2026100003-3"
"08/05/2025","SO","SO IRONBRIDGE TOOLING REF320672","-7650.56","LEDGERBENCH DEMO LTD","74-30-71 79855615","acct-2026100003-3"
"09/05/2025","DD","DD JUNIPER PAYROLL REF244932","-13468.65","LEDGERBENCH DEMO LTD","74-30-71 79855615","acct-2026100003-3"
"10/05/2025","SO","SO DELPHINIUM MEDIA REF318018","-10287.03","LEDGERBENCH DEMO LTD","74-30-71 79855615","acct-2026100003-3"
"10/05/2025","FPO","FPO OSPREY LEASING REF294938","-4533.69","LEDGERBENCH DEMO LTD","74-30-71 79855615","acct-2026100003-3"
"12/05/2025","CR","CR OSPREY LEASING REF760756","1179.91","LEDGERBENCH DEMO LTD","74-30-71 79855615","acct-2026100003-3"
"Account Number","Posting Date","Transaction Type","Account Name","Amount","Ledger Ref","Narrative"
"74-30-71 79855615","12/05/2025","DD","LEDGERBENCH DEMO LTD","-2889.35","acct-2026100003-3","DD DELPHINIUM MEDIA REF287062"
"74-30-71 79855615","14/05/2025","FPI","LEDGERBENCH DEMO LTD","8053.10","acct-2026100003-3","FPI IRONBRIDGE TOOLING REF490116"
"74-30-71 79855615","14/05/2025","CR","LEDGERBENCH DEMO LTD","10995.87","acct-2026100003-3","CR EASTFIELD ENERGY REF138926"
"74-30-71 79855615","15/05/2025","FPI","LEDGERBENCH DEMO LTD","2334.98","acct-2026100003-3","FPI DELPHINIUM MEDIA REF876828"
"74-30-71 79855615","15/05/2025","CR","LEDGERBENCH DEMO LTD","14449.94","acct-2026100003-3","CR OSPREY LEASING REF738241"
"74-30-71 79855615","16/05/2025","CR","LEDGERBENCH DEMO LTD","11215.07","acct-2026100003-3","CR HARBOURLIGHT IT REF703150"
"74-30-71 79855615","17/05/2025","DD","LEDGERBENCH DEMO LTD","-13629.96","acct-2026100003-3","DD JUNIPER PAYROLL REF394821"
"74-30-71 79855615","18/05/2025","FPI","LEDGERBENCH DEMO LTD","10772.63","acct-2026100003-3","FPI MOORLAND TELECOM REF183479"
"74-30-71 79855615","19/05/2025","SO","LEDGERBENCH DEMO LTD","-8554.11","acct-2026100003-3","SO NORTHGATE CLEANING REF808200"
"74-30-71 79855615","21/05/2025","TFR","LEDGERBENCH DEMO LTD","11597.12","acct-2026100003-3","TFR JUNIPER PAYROLL REF816508"
"74-30-71 79855615","24/05/2025","FPO","LEDGERBENCH DEMO LTD","-11257.94","acct-2026100003-3","FPO IRONBRIDGE TOOLING REF546570"
"74-30-71 79855615","25/05/2025","TFR","LEDGERBENCH DEMO LTD","13764.83","acct-2026100003-3","TFR OSPREY LEASING REF454290"
"74-30-71 79855615","28/05/2025","BP","LEDGERBENCH DEMO LTD","-2937.15","acct-2026100003-3","BP JUNIPER PAYROLL REF990796"
"74-30-71 79855615","30/05/2025","BP","LEDGERBENCH DEMO LTD","-10605.96","acct-2026100003-3","BP NORTHGATE CLEANING REF398558"
 

Ground truth

Public here because the benchmark is open; it is the model's prompt path that can never reach these values, not human readers.

AccountClosing balance
acct-2026100003-1£26,791.55
acct-2026100003-2£43,913.12
acct-2026100003-3£88,858.72

Recorded runs on this case

  • exactSingle prompt2 steps · 11,731 tokens · Confidence high

    Model claude-sonnet-4-6Measured 2026-08-20Bench v1.0Run 11eaa71b2d02Commit 3d38c30

  • exactAgent with tools20 steps · 44,032 tokens · Confidence high

    Model claude-sonnet-4-6Measured 2026-08-20Bench v1.0Run f140aa1e57f4Commit 384d72d

Step through any of these on the demo page.