---
library_name: portallib
license: apache-2.0
base_model: Qwen/Qwen3-8B
datasets:
- RampPublic/portallib-tasks
tags:
- portal
- hypernetwork
- lora
- peft
- multiple-choice
---

# PorTAL refit for Qwen3-8B

This is a native [PorTAL](https://github.com/ramp-public/portallib) artifact refitted onto
`Qwen/Qwen3-8B`. Its 14-task latent table and canonical LoRA-generating core were learned jointly
from Qwen3-1.7B and Qwen3-4B and frozen during refitting. Only a fresh Qwen3-8B alignment was
trained. The artifact generates rank-8 LoRA factors for the query and value projections of every
decoder layer.

## Evaluation

One seed was evaluated on the complete 14-task validation suite using continuation log-probability
divided by character length (`acc_norm`). Gold continuation token-mean NLL was tracked separately
for checkpoint selection.

| Model | Macro `acc_norm` |
|---|---:|
| Frozen Qwen3-8B | 0.6681 |
| PorTAL-adapted | 0.7767 |
| Absolute lift | +0.1086 |

These are research benchmark results for this exact artifact and evaluation recipe, not a general
performance guarantee.

## Refit recipe

- Base: `Qwen/Qwen3-8B` at `b968826d9c46dd6066d109eabc6255188de91218`
- Frozen source carrier: `RampPublic/portal-qwen3-4b`
- Dataset: `RampPublic/portallib-tasks` at
  `ffc3c0e44f529bf64a5ae62ed5db090952db97ea`
- Refit data: deterministic seeded sample of up to 1,000 examples per task from the complete
  training pool
- Optimization: 5 epochs, batch size 4, alignment LR `1e-3`, linear decay with 10% warmup, seed 0
- Trainable parameters: target-base alignment only; task latents and canonical core remain frozen
- Checkpoint: maximum macro validation `acc_norm`, with lower gold NLL as the tie-breaker
- Architecture: q/v targets, rank 8, alpha 16, task latent 256, layer embedding 32, hidden 512,
  canonical width 1024

## Usage

```python
from portallib import PortalModel

portal = PortalModel.from_pretrained(
    "RampPublic/portal-qwen3-8b",
    revision="v0.2.0",
)
portal.export_peft("rte", "./portal-rte-qwen3-8b")
```

See the [release recipe](https://github.com/ramp-public/portallib/blob/main/REPRODUCING.md) for the
full task list, evaluation definition, and refitting procedure. The artifact is Apache-2.0; the
benchmark dataset contains components under multiple upstream licenses documented on its dataset
card.