Docs
OfficialJev Practices & Patterns
Jev 1.13 Jaggedness
A maintained list of jev-1.13's known failure modes, literal reading, unreliable counting, dates as text, indirection, context rot, and contradictory criteria, each with a guardrail.
- Format
- Docs
- Reading time
- 2 min
- Published by
- TypeSafe AI
- Last verified
- Sep 2026
Highlights
- Counting is unreliable and error grows with size: count in code, one question per item.
- Dates are read as text; extract components with Choice and compare or order them in code.
- Score levels are weak at numeric calibration, so don't interpolate an exact magnitude between levels.
- Structural invariants do not hold: complementary probabilities can sum past 1 and the same question can disagree across primitives.
Quickstart
python
result = client.system_one(
{"items": items},
{
f"item_{i}": Noul(instructions=f"Is `items[{i}]` the name of a fruit?")
for i in range(len(items))
},
)
count = sum(result.nouls[f"item_{i}"].noul > YES for i in range(len(items)))Watch out
The list is reviewed periodically rather than continuously, so treat it as a starting checklist and add the failure modes your own evals find.
Questions about this entry
Short answers drawn from the linked project or page.
Don't ask for a total. The page lists counting as unreliable with error growing with size, and its guardrail is one Noul per item with the sum done in code — the snippet counts per-item booleans rather than trusting a single count answer.
Do it in code. The page says dates are read as text and extracted as components with Choice for comparison outside the model, and warns that Score levels are weak at numeric calibration, so don't interpolate an exact magnitude between levels.
jev-1.13, last reviewed 2026-10-02. Many of the listed edges are expected to be fixed in later versions.
State the exact condition in the instructions and put boundary cases in the criteria, because jev-1.13 answers the words you wrote rather than the intent behind them.
Treat it as a Choice over enumerated parts such as month, day and year, include an explicit not-stated option, then assemble and compare the real date in code.
No. The page notes structural invariants do not hold: complementary probabilities can sum past 1, and the same question can disagree across primitives.
Filter first and send only what the question needs, because irrelevant context distracts the model.
More Jev resources like this
CLI that fits a per-question confidence threshold to a target accuracy on your own labeled data, verifies it on a held-out split, estimates how much traffic still needs an LLM, and re-checks locked thresholds in CI. It publishes no Jev results of its own.
Jev Practices & Patterns#community#python#calibration
Community
abhixhek/jevcalPython
abhixhek/jevcalPythonAn independent measurement spanning 5,477 test rows and 34.1M input tokens, comparing one direct Jev question per row with 12-14 scored dimensions fitted to local labels.
Jev Practices & PatternsArticle#community#benchmarks#evaluation
Community
How to shape what Jev evaluates: a string, a named object, or an array of messages. Every question in a request sees the same state and is evaluated independently.
Jev Practices & PatternsDocs#official#docs#state
Official
Rate content against 2 to 10 ordered levels. Returns a probability-weighted score that can land between levels, per-level probabilities, and confidence, with the arithmetic left to code.
Jev Practices & PatternsDocs#official#primitives#score
Official
Classify intent with a Choice and complexity with a Score, then send each branch to deterministic code, a specialist model, or a person, with a 0.5 intent-confidence floor.
Jev Practices & PatternsDocs#official#patterns#routing
Official
A self-hosted implementation of TypeSafe's Jev System One API powered by the 400M-parameter GLiFormer encoder: it serves choice, score, and noul and drops into the official typesafe-sdk via TYPESAFE_BASE_URL, but trails Jev on reasoning-heavy tasks.
Jev Practices & Patterns#community#python#open-models
Community
logan-markewich/jeffPython
logan-markewich/jeffPythonFrom the Jev guides
Original write-ups on what this entry is for.
From the Jev AI model community
Posts from builders shipping with Jev right now.