AnyJev
A training-free layer that turns an open LLM into a decision model. It reads typed choice, yes/no and score answers from one prefill, removes option-order bias with no labels and, from a few hundred labels, calibrates or fits a closed-form head. Weights stay untouched.
Every decision carries the level that produced it. L0 needs no labels: it asks once per rotation of the options and divides out a label prior (on Qwen3-8B, BANKING77 20-way, answer flips under reversed options fall from 0.230 to 0.073). Since 0.2.0 it can stop once the leader is far enough ahead, calibrated without labels to match the full cycle: 7.2 rotations instead of 18 at a certified 1% disagreement, 2.2x the decisions per second on vLLM. L1 adds a temperature from 100 to 500 labels; L2 solves a closed-form head per question on a hidden state partway down, from 100 to 300 labels. This entry is AnyJev on Qwen3-8B: on LocalLLaMA/typed-decisions L2 scores 77.1% with a calibration error of 0.034 at block 24 of 36 (Jev's published figure: 72.7%). Heads for five Qwen3 models, 1.7B to 32B (73.0% to 79.9%), ship in the GitHub repository rather than on Hugging Face, and the Qwen weights come from Qwen, so there are no files to pull here. vLLM serves every level, L2 included. A head serves one question on one model, and the letter readout takes at most 26 options. Code and heads are Apache-2.0. By Jiamu Zhang, Tianze Yang and Liang Wu (Nokia) with Yucheng Shi (Tencent Hunyuan).
What it decides
- choice — picks one option from a set
- score — places the input on an ordered scale
- noul — answers a yes/no question with one calibrated probability
- classify — assigns a category from a fixed taxonomy
- route — sends the input to one of several destinations
At a glance
| Parameters | 8B |
| Base model | qwen/qwen3-8b |
| Maker | Nokia |
| Released | 2026-09-21 |
| License | apache-2.0 |
| Reported accuracy | 77.1% |
| Reported latency | 54.6 ms for one L0 decision (K=4, 110-token state), 42.8 ms per decision in batches of 32, on one H100 NVL with transformers; L2 at block 24 costs 0.68x a full forward |