Bigger Context Home windows Don’t Repair RAG — So I Constructed a System That Does

TL;DR

• I constructed a dataset Q&A system and trusted a RAG reply that was lower than half-correct.
• I measured this throughout 7 question varieties and 5 context sizes on 100,000 rows.
• queries away from RAG totally.

I Trusted the Incorrect Quantity

Final month I used to be heads-down constructing a brand new characteristic for EmiTechLogic. Learners can now add their very own messy CSV information and ask questions in plain English about their information. Sounded excellent for RAG, so I went all in — embeddings, retrieval, nice-looking responses.

The primary few demos appeared wonderful. Clear tables, assured numbers, skilled formatting. I really began trusting the system in our inner testing.

Then I picked one quantity to double-check.

Actual grocery spend within the dataset: $1,140,033.24.  

The mannequin gave me a wonderful breakdown by class. It appeared legit. I added up the numbers it returned.

It was lower than half.

I sat there staring on the display screen considering “this will’t be proper.” So I did what any engineer would do. I elevated the context window. 4k… 16k… 32k… 128k tokens. Every time the reply acquired longer, extra detailed, and extra confidently flawed.

That’s when it lastly clicked. This wasn’t a retrieval challenge. I used to be asking a retrieval system to carry out heavy computation on information it had solely partially seen. And as a substitute of claiming it was not sure or lacking data, the mannequin was producing polished, structured solutions that appeared right.

Why RAG Can not Combination

The RAG pipeline doesn’t really perceive structured information. All it does is take every CSV row and flatten it into plain textual content. That’s it. To the mannequin, a row seems one thing like this:

"2019-01-01 grocery_pos 107.23 F NC Jennifer Banks ..."

For a question like “What’s the complete spend by class?”, the RAG pipeline does this:

1. Tokenise: ["total", "spend", "category"]
2. Rating all 100,000 rows by key phrase overlap
3. Return the top-N rows as serialised plain textual content
4. Ask the LLM to sum and group from that textual content

Step 4 is the place the system fails. The LLM isn’t operating a SUM. It’s pattern-matching numbers from a textual content blob and producing a response that mimics an aggregation.

Fashions battle with numerical precision at scale [1], however the true challenge is the presentation. The mannequin provides you an in depth breakdown throughout all classes. It is a basic lure. The output seems skilled. It mimics the construction of an actual report so properly that your mind assumes the content material is legitimate. You haven’t any approach to confirm that 92% of your information is lacking.

RAG is a retrieval instrument. It isn’t a calculation engine. Retrieval finds related fragments. Computation requires a full dataset scan. Whenever you use RAG for math, you get a flawed reply that appears authoritative. That distinction is important. A partial reply indicators that information is lacking. A whole-looking flawed reply simply indicators false confidence.

Full code: https://github.com/Emmimal/context-window-engine/

The Benchmark: Two Pipelines, Identical Question

To measure this exactly, I constructed a benchmark that runs two pipelines facet by facet for each question.

The primary pipeline is a RAG simulation. It fashions what a naive vector pipeline passes to an LLM at 5 context sizes. I examined 5 context sizes, starting from 5 rows as much as 8,000. That scales from 325 tokens to 500,000. For every measurement, I tracked three metrics: how a lot information the LLM sees, what sum it computes from that particular slice, and whether or not a reader may really spot the error.

The second pipeline is a semantic engine that executes the identical question as a deterministic full-scan over all 100,000 rows and returns the precise right reply.

The simulation doesn’t reproduce actual LLM outputs. What it preserves is the important thing structural property: a partial slice of information fed right into a system that returns a full-form reply. That’s the property that causes the issue, and that’s what the benchmark measures.

I selected seven question varieties to cowl each aggregation sample a structured information system is more likely to encounter:

These queries usually are not distinctive or advanced. They’re the usual questions any analyst asks when taking a look at a brand new dataset. That’s precisely why this failure is so important.

Error Observability Collapse

Right here is the total benchmark output for the question that began all of this. I’m exhibiting it in full as a result of the numbers make the issue unimaginable to dismiss.

GROUND TRUTH (Semantic Engine)
SUM(amt) GROUP BY class → 14 teams
  #1  grocery_pos               1,140,033.24
  #2  shopping_net                773,527.93
  #3  shopping_pos                725,766.14
  #4  gas_transport               648,804.24
  #5  house                        556,526.53
Latency: 100.47ms | Rows scanned: 100,000

RAG SIMULATION — what the LLM receives at every context measurement

Context               Rows   Protection    Partial sum  Error detectable?
tiny   (~325 tokens)     5   0.0050%         197.73  EASY
small  (~3K tokens)     50   0.0500%       2,003.56  MODERATE
medium (~32K tokens)   500   0.5000%      31,023.21  HARD
massive  (~130K tokens) 2,000  2.0000%     140,093.16  VERY HARD
xlarge (~520K tokens) 8,000  8.0000%     569,368.22  NEAR IMPOSSIBLE

I stared at these outcomes for some time. Essentially the most troubling half wasn’t simply that the solutions have been flawed, it was how a lot tougher the errors turned to identify because the context window grew.

At 8,000 rows the error was nonetheless over 50%, but the response appeared like knowledgeable report. You’d must manually confirm the numbers to note one thing was off. That’s what I began calling Error Observability Collapse. The extra context I gave the mannequin, the extra convincing — however no more correct — the output turned.

The “Partial sum” column reveals the entire if the LLM added each quantity worth within the rows it really retrieved. The “Error detectable?” column scores how probably a human reader is to identify a mistake.

With 5 rows, the partial sum is 197.73. The right complete is 1,140,033.24. It’s apparent. The output is brief, the numbers are flawed, and the lacking information is obvious. The error is immediate.

At 8,000 rows, the partial sum hits 569,368.22. The LLM has now seen all 14 classes. It generates a 1,500-word report with particular figures and assured language. The error is 50%, however it’s hidden inside authoritative, well-structured prose. With out an exterior reference, a reader has no approach to catch it.

That is the sample that held throughout all seven queries:

I referred to as this Error Observability Collapse. As context grows, confidence scales with it. Correctness doesn’t.

The failure modes are uneven, which makes them harmful:

A flawed RAG reply seems right. It’s formatted, particular, and assured. A failed computation throws an express error. It’s seen.

One failure is silent. The opposite is loud. As context home windows attain tens of millions of tokens, the silent failure turns into tougher to detect [4]. The system doesn’t get safer because it scales. It simply will get extra convincing.

The Semantic Engine: Proof That the Right Reply Is Quick

Earlier than I totally understood the issue, I had already thrown collectively a easy semantic engine out of frustration. I simply wished the right reply no less than as soon as.

The method turned out to be easy: parse the question into correct operations and run a single cross over the complete dataset. No embeddings, no retrieval, no guessing.

Right here’s what that appears like in observe:

The logic is straightforward. Take a question like “What’s the complete spend by class?”. The engine maps this to a direct operation: SUM(amt) GROUP BY class. It processes the total 100,000-row set in a single cross. It accumulates grouped totals. There is no such thing as a retrieval. No inference. No partial scanning. It visits each row as soon as and returns the precise outcome.

This proves that the right reply isn’t costly. Benchmark queries completed below 200ms. Pattern measurement: 100,000 rows. Aggregation is trivial. The failure occurs while you route these queries to a system constructed to misconceive them.

from context_window_engine import compute_ground_truth, load_csv

rows = load_csv("information/credit_card_transactions.csv", max_rows=100_000)

gt = compute_ground_truth(
    query_label = "complete by class",
    rows        = rows,
    agg_func    = "sum",
    agg_col     = "amt",
    group_col   = "class",
)
# gt.reply     → [(grocery_pos, 1140033.24), (shopping_net, 773527.93), ...]
# gt.latency_ms → 100.47

Engine helps SUM, AVG, COUNT, MIN, MAX. Handles categorical and numeric filters. Contains GROUP BY and ratios. Zero exterior dependencies. Each operation runs as a deterministic operate over the total listing.

The engine itself isn’t the product. It’s the proof: the right reply is reachable below a second. No inference required. The true problem is routing queries there reliably.

The Repair Is Not Higher Retrieval

Cease attempting to enhance retrieval. If a question wants 100% of the info, an 8% pattern fails. The repair is eradicating retrieval from the loop.

We want a classification layer. It sits earlier than the pipeline and makes one binary name: computation or lookup?

The distinction is obvious. “Complete spend by class” calls for a full scan. “Discover transactions from Jennifer Banks” is a straightforward lookup. Normal RAG forces each down the identical path. That’s the design flaw.

A QueryRouter fixes this. It inspects each incoming question and routes it to the right path earlier than a single retrieval begins.

The classifier makes use of three sign tiers, prioritized. Tier 1: aggregation verbs—complete, what number of, common, lowest, proportion. These demand full-dataset computation. Tier 2: numeric comparability—better than 500, above $1,000, no less than. These indicate filter-then-aggregate, unimaginable for RAG. Tier 3: retrieval indicators—discover, present me, listing, fetch. These point out lookups the place semantic similarity works.

Default to COMPUTATION if no tier matches. That is deliberate. Failure modes are uneven: a flawed RAG reply on an aggregation is silently flawed. A computation engine that may’t parse a question throws an error. When doubtful, fail loudly.

from query_router import QueryRouter

router = QueryRouter(rows)

outcome = router.route("What's the complete spend by class?")
# outcome.routed_to     → "COMPUTATION"
# outcome.reply.reply → [(grocery_pos, 1140033.24), ...]
# outcome.total_latency → ~250ms — classify + execute mixed

outcome = router.route("Discover transactions from Jennifer Banks")
# outcome.routed_to     → "RETRIEVAL"
# outcome.reply.secure   → True — RAG is acceptable

Routing the Full Benchmark

I ran 9 queries by means of the router to confirm efficiency throughout each varieties: seven aggregation queries destined for the semantic engine, and two lookup queries for RAG.

Each route was right. The seven aggregation queries hit the full-scan engine and returned actual outcomes. The 2 lookup queries appropriately triggered the RAG path. Take a look at the output: excessive confidence scores, right sample matching, and latency below 130ms—even with the 100,000-row scan.

[1] ✓  COMPUTATION   "What's the complete spend by class?"
     Tier 1 | matched='complete' | confidence=0.97
     #1 grocery_pos      1,140,033.24  (102.57ms | 100,000 rows | actual)

[2] ✓  COMPUTATION   "Which class has the very best common transaction quantity?"
     Tier 1 | matched='highest' | confidence=0.97
                               71.91  (119.47ms | 100,000 rows | actual)

[3] ✓  COMPUTATION   "What's the complete quantity spent on grocery_pos?"
     Tier 1 | matched='complete' | confidence=0.97
                        1,140,033.24  (49.96ms  | 100,000 rows | actual)

[4] ✓  COMPUTATION   "What number of transactions have been made by feminine prospects?"
     Tier 1 | matched='What number of' | confidence=0.97
                           54,641.00  (90.45ms  | 100,000 rows | actual)

[5] ✓  COMPUTATION   "What's the complete spend the place quantity is larger than 500?"
     Tier 1 | matched='complete' | confidence=0.97
                        1,274,269.60  (91.65ms  | 100,000 rows | actual)

[6] ✓  COMPUTATION   "Which state has the bottom complete spending?"
     Tier 1 | matched='lowest' | confidence=0.97
     lowest RI               2,125.60  (109.05ms | 100,000 rows | actual)

[7] ✓  COMPUTATION   "What proportion of transactions are fraudulent?"
     Tier 1 | matched='proportion' | confidence=0.97
                              0.9900%  (87.35ms  | 100,000 rows | actual)

[8] ✓  RETRIEVAL     "Discover transactions from Jennifer Banks"
     Tier 3 | matched='Discover' | confidence=0.85
     RAG is acceptable — no aggregation required

[9] ✓  RETRIEVAL     "Present me a pattern transaction from Texas"
     Tier 3 | matched='Present me' | confidence=0.85
     RAG is acceptable — no aggregation required

Routing accuracy: 9/9

9/9 right. Error Observability Collapse is unimaginable if aggregation queries by no means attain RAG.

The Check Suite

The benchmark verifies 9 particular queries. The take a look at suite ensures reliability throughout a broader vary: edge instances, malformed inputs, lacking information, and customary manufacturing failure factors.

The engine suite has 87 checks throughout 10 courses. It covers float parsing with greenback indicators, commas, and scientific notation; all 5 aggregation capabilities below regular circumstances and with empty inputs; all 5 numeric filter operators; full GROUP BY aggregation with categorical and numeric filters mixed; RAG simulation protection metrics at every context measurement; and edge instances together with empty datasets, rows with lacking column values, and single-row inputs.

The router suite has 72 checks throughout 5 courses. It covers all three tier patterns, together with edge instances like all-caps queries and really lengthy queries; pure language to typed operation parsing for each supported question type; routing and execution correctness in opposition to all seven benchmark queries; and a distinction suite that verifies router solutions match impartial ground-truth computation — guaranteeing the router doesn’t introduce any deviation from the engine’s personal output.

Run the engine checks by typing python area -m area unittest area test_engine area -v. This executes the 87 checks within the suite.

Run the router checks by typing python area -m area unittest area test_router area -v. This executes the 72 checks within the suite.

All 159 cross on Python 3.9+ with zero exterior dependencies.

Trustworthy Limitations

This answer isn’t excellent. It solely works on single CSV information proper now. Actual manufacturing datasets are normally messy with a number of tables that want becoming a member of — I intentionally stored the scope small as a result of I wished one thing that truly labored end-to-end first.

The router can also be nonetheless fairly fundamental (regex-based). I attempted a small LLM-based classifier early on but it surely was inconsistent and added latency, so I went again to the straightforward method. Typically the boring answer wins.

I additionally simulated the RAG responses as a substitute of hitting actual APIs for the benchmark. The patterns maintain up, however your mileage with GPT-4o or Claude 3.5 will range barely.

CSV format required. The engine masses structured information immediately from CSV information. Database connections, Parquet information, and different tabular codecs usually are not supported presently.

What This Adjustments

Including a routing layer prices virtually nothing. Classifying a question in opposition to 65 regex patterns takes simply microseconds. The semantic engine provides lower than 200ms to scan a 100,000 row dataset. The entire overhead is smaller than a single embedding name.

What you get in return is a deterministic reply for each aggregation question. Each complete, each depend, and each proportion now comes from a full scan as a substitute of a assured approximation based mostly on 8 p.c of the info. RAG retains dealing with what it’s really good at: retrieving particular data, surfacing related passages, and answering lookup questions the place semantic similarity is the correct instrument for the job.

RAG isn’t damaged. It’s simply being requested to compute, and it can’t try this.
The harmful half isn’t that it fails. It’s that it fails convincingly. And no quantity of context adjustments that.

You’ll be able to attempt typing it out like this:

To begin, clone the repository utilizing git clone adopted by the URL https://github.com/Emmimal/context-window-engine/. As soon as that finishes, transfer into the listing by typing cd context-window-engine. Lastly, launch the mission by operating python demo.py in your terminal.

References

[1] Levy, M., Jacoby, A., & Goldberg, Y. (2024). Identical activity, extra tokens: The influence of enter size on the reasoning efficiency of huge language fashions. In Proceedings of the 62nd Annual Assembly of the Affiliation for Computational Linguistics (Quantity 1: Lengthy Papers), pages 15339–15353, Bangkok, Thailand. Affiliation for Computational Linguistics.
https://doi.org/10.18653/v1/2024.acl-long.818

[2] Lewis, P., Perez, E., Piktus, A., Petroni, F., Karpukhin, V., Goyal, N., Küttler, H., Lewis, M., Yih, W.-t., Rocktäschel, T., Riedel, S., & Kiela, D. (2020). Retrieval-augmented era for knowledge-intensive NLP duties. Advances in Neural Info Processing Methods, 33, 9459–9474. https://doi.org/10.48550/arXiv.2005.11401

[3] Gao, Y., Xiong, Y., Gao, X., Jia, Okay., Pan, J., Bi, Y., Dai, Y., Solar, J.,
Guo, Q., Wang, M., & Wang, H. (2023). Retrieval-augmented era for giant language fashions: A survey. arXiv preprint arXiv:2312.10997.
https://doi.org/10.48550/arXiv.2312.10997

[4] Liu, N. F., Lin, Okay., Hewitt, J., Paranjape, A., Bevilacqua, M., Petroni, F., & Liang, P. (2023). Misplaced within the center: How language fashions use lengthy contexts. Transactions of the Affiliation for Computational Linguistics, 12, 157–173. https://doi.org/10.1162/tacl_a_00638

[5] Koshorek, O., Granot, N., Alloni, A., Admati, S., Hendel, R., Weiss, I., Arazi, A., Cohen, S.-N., & Belinkov, Y. (2025). Structured RAG for answering aggregative questions. arXiv preprint arXiv:2511.08505.
https://doi.org/10.48550/arXiv.2511.08505

Disclosure

All benchmark numbers are from precise runs on Python 3.12.6, Home windows 11, CPU solely, no GPU. The benchmark makes use of the Credit score Card Transactions Fraud Detection dataset (Kartik Gajjar, Kaggle, 2020), an artificial dataset generated utilizing the Sparkov transaction simulator created by Brandon Harris, licensed CC0 (Public Area), obtainable at kaggle.com/datasets/kartik2112/fraud-detection. The RAG baseline simulates retrieval and fashions confidence indicators — no actual LLM API calls are made. No exterior API keys are required to breed any outcome on this article. All code described right here was written and examined by me.