Not every AI call needs to generate text: ej, an 11 MB model for typed decisions

A developer has released ej, an 11 MB open model that answers typed decision questions — choice, yes/no, and ordered score — by returning one probability distribution per question in a single forward pass without generating any tokens. The model, installable via pip as ejai (version 0.0.1), runs on CPU and lets callers threshold on probabilities rather than parse generated JSON, positioning it as a lighter alternative to prompting an LLM or running per-pair NLI zero-shot classifiers.

A lot of "AI" in production pipelines isn't writing anything. It's deciding. Which queue does this ticket go to? Does this message need a human? How urgent is this request: low, medium or high? The answer is one item from a list you already know. Yet the usual way to get it is to send the text to a language model, ask it to reply in JSON, parse the reply, retry when the JSON is broken, and hope the word "high" means the same thing every time. That works, but it's an odd fit. You pay for generation you don't need, you get text where you wanted a number, and you have no clean way to say "only automate this when the model is at least 90% sure." I built ej to do that one job differently. It's an open model that takes a piece of text plus a set of typed questions, and returns one probability distribution per question, in one forward pass, without generating a single token . The whole model is one 11 MB file that runs on a CPU. pip install ejai the import name is ej This is version 0.0.1. Below is what it does, how I measured it, and, just as important, where it is weak. You give ej a state free text, or a JSON object written as text and any number of questions. Each question has one of three types: | Type | You give | You get | |---|---|---| | choice | instructions + a list of option texts, chosen when you call it | a probability for each option | | noul | a yes/no statement | P false , P true | | score | instructions + ordered levels | a distribution over the levels | The options are plain text you write at call time. There is no fixed label set baked in at training. python import ej model = ej.load "5ak3t/ej", revision="v0.0.1" downloads model.ejpack and checks its SHA-256 record = { "state": '{"customer tier": "gold", "message": "The blender arrived with a cracked jug. Replace it before Friday."}', "questions": { "route": {"type": "choice", "instructions": "Which team should handle this request?", "options": {"key": "returns", "text": "Returns and replacements for damaged or wrong items"}, {"key": "billing", "text": "Billing, invoices and payment problems"} }, "needs human": {"type": "noul", "instructions": "The customer is upset enough that a human agent should reply.", "options": ej.NOUL OPTIONS}, "urgency": {"type": "score", "instructions": "How urgent is this request?", "options": {"key": "0", "text": "Low: can wait a week"}, {"key": "1", "text": "Medium: answer within two days"}, {"key": "2", "text": "High: answer today"} }, }, } probs, = model.predict record print probs "route" p returns, p billing , sums to 1 print probs "needs human" p false, p true print probs "urgency" p low, p medium, p high All three questions are answered in the same pass. Because the output is a distribution, the routing logic in your code is a threshold, not a string parser: if max probs "route" = threshold: auto route record, probs "route" else: send to human record Pick that threshold on labelled examples of your workflow. More on why below. There are three common ways to make this kind of decision today. ej is a different trade-off from each, not a strict upgrade. Prompting an LLM. It's flexible and strong on new tasks. But it generates text you have to parse, it needs either an API call cost, latency, your data leaving your servers or a local model that is hundreds of MB to several GB, and its reply isn't a probability you can threshold. ej skips generation entirely and is small enough to ship inside a service. NLI zero-shot classifiers the "zero-shot-classification" pipeline . These are the closest relatives. They usually run one forward pass per text, option pair, so cost grows with the number of options, and each question type needs its own setup. ej reads the text once and answers choice, yes/no and score questions together. To be clear, though, a good NLI model beats ej on workflows ej has never seen numbers below . Fine-tuning your own classifier BERT, SetFit and friends . It's accurate on its own task, but the label set is fixed at training, and every change means new labelled data and a retrain. With ej the options are input. It also has an adapt method that takes a handful of labelled records: adapted = model.adapt examples=labelled :8 per-option offsets; the weights don't change probs = adapted.predict new records Briefly a technical report with the full method is coming : e5-small-v2 , quantised to model.ejpack : On the test box a shared 4-core Xeon, one thread , a warm call takes roughly 200 to 360 ms per record , depending on the run and the inputs, and the process peaks at about 558 MB of RAM. Most of that is torch itself: the model's own tensors peak at 128 MB, or 35 MB with ej.load ..., low memory=True , which is about 1.5x slower and gives identical predictions. I tried hard to make these numbers trustworthy rather than flattering: | Test suite | ej 0.0.1 accuracy 95% CI | |---|---| | Typed Decisions, workflows seen in training | .721 .695, .746 | | MASSIVE intents leak-free split | .832 .808, .857 | | Support tickets in-house | .712 .679, .746 | | Support tickets, new writing styles | .683 .639, .726 | | 105 workflows never seen in training | .419 .379, .458 | | One held-out Typed Decisions workflow also used during model selection | .422 .388, .452 | How that compares with the six rivals: A few more things you should know before using it: ej makes sense if all of these are true: If your task is brand-new and you have no labelled data at all, a prompted LLM or an NLI zero-shot model will probably serve you better today. I'd rather say that here than have you find out in production. pip install ejai python import ej model = ej.load "5ak3t/ej", revision="v0.0.1" The repository also contains the training code python -m ej.train , the evaluator that produced every number above python -m ej.eval , and the benchmark runner with the rival adapters, so you can check my numbers or train your own pack. If you run it on your own workflow, I'd love to hear what accuracy you get, good or bad. Results on real workflows are exactly what version 0.0.2 needs.