# ProbeLoop Laya stated limits decision model by submsk

> ProbeLoop Laya stated limits finds explicit knowledge gaps in Italian practice answers. Its provisional ONNX classifier uses synthetic training data.

Updated 2026-11-12 (Australia/Adelaide)

## Quick answers

### What is ProbeLoop Laya stated limits by submsk?

[ProbeLoop Laya stated limits](<https://decisionmodels.ai/models/probeloop-laya>): A provisional Italian sentence classifier fine-tuned from [Laya](<https://decisionmodels.ai/models/laya>) multilingual.

[ProbeLoop Laya stated limits facts](<https://decisionmodels.ai/models/probeloop-laya>)

[Source](<https://huggingface.co/submsk/probeloop-laya-stated-limits>)

### Who makes ProbeLoop Laya stated limits?

[ProbeLoop Laya stated limits](<https://decisionmodels.ai/models/probeloop-laya>): submsk

[ProbeLoop Laya stated limits facts](<https://decisionmodels.ai/models/probeloop-laya>)

[Source](<https://huggingface.co/submsk/probeloop-laya-stated-limits>)

By submsk · Hugging Face weight release: 2026-11-10 · OPEN

A provisional Italian sentence classifier fine-tuned from [Laya](<https://decisionmodels.ai/models/laya>) multilingual. It finds explicit statements of missing knowledge or experience in interview-practice answers. This is a fixed two-class encoder and head, separate from [Laya](<https://decisionmodels.ai/models/laya>)’s general typed-question heads; it does not assess competence or choose interview moves.

## Facts

- Size: 306,941,186 parameters; full float32 weights and a 308 MB dynamic-int8 ONNX export
- Inputs: One answer sentence paired with the last 300 characters of the interviewer’s question; 192-token truncation
- Runtime: ONNX Runtime, Tokenizers and NumPy on CPU; no [Laya](<https://decisionmodels.ai/models/laya>) SDK required
- Training: 2,549 synthetic Italian sentence pairs, 362 positive; final checkpoint trained on all rows
- Calibration: Threshold 0.5; no calibrated confidence guarantee or real-interview acceptance test
- Licence: Apache-2.0 fine-tuning and exports; retain [Laya](<https://decisionmodels.ai/models/laya>) and mmBERT upstream attribution

## Claims and sources

- **VENDOR-REPORTED**: The maker reports 288/294 positive answers found and 8/312 negative answers flagged in a synthetic leave-one-profession-out pilot. Those results concern held-out-group models, not an unseen test of the final checkpoint trained on all rows. [source](<https://huggingface.co/submsk/probeloop-laya-stated-limits>)

## Links

- [Source: Card and weights](<https://huggingface.co/submsk/probeloop-laya-stated-limits>)
- [Source: Licence](<https://huggingface.co/submsk/probeloop-laya-stated-limits/blob/main/LICENSE>)
- [Source: Checksums and provenance](<https://huggingface.co/submsk/probeloop-laya-stated-limits/blob/main/manifest.json>)
- [Source: Pilot and integration record](<https://github.com/subnetMusk/probeloop/blob/main/docs/research/laya-limits-pilot.md>)

## Related projects

### ProbeLoop stated-limit reader

[ProbeLoop stated-limit reader](<https://decisionmodels.ai/projects#project-probeloop>)

[Source: subnetMusk/probeloop](<https://github.com/subnetMusk/probeloop>)

Uses a fine-tuned [Laya](<https://decisionmodels.ai/models/laya>) ONNX classifier to find explicit knowledge or experience limits in Italian practice answers. A recorded synthetic HTTP/WebSocket turn uses the actual classifier and scripted LLMs. Interview decisions and evaluation remain separate; real-speech quality and physical Mac performance are still open.

- Creator: subnetMusk

- Stars: 0

- Licence: Apache-2.0

- Model: [ProbeLoop Laya stated limits](<https://decisionmodels.ai/models/probeloop-laya>)

- Tags: local, interviews

Author-run records; not independently reproduced

- [Source: Licence](<https://github.com/subnetMusk/probeloop/blob/main/LICENSE>)
- [Source: Pilot and integration run](<https://github.com/subnetMusk/probeloop/blob/main/docs/research/laya-limits-pilot.md>)
- [Source: Current acceptance limits](<https://github.com/subnetMusk/probeloop/blob/main/docs/status.md>)