<!-- agent-source-envelope -->
# Artifact: Qwen3-4B Multibackend Distilled | PTLLM

> A public failed attempt that retained 100% file-ops depth but drove the recorded breadth score down to 31%.

Canonical: https://posttrainllm.com/artifacts/qwen3-4b-multibackend-distilled
<!-- /agent-source-envelope -->

Negative-transfer case study

Report artifact

2026-07-03

Adding multibackend teacher data did not create a broader agent. It preserved the saturated file-ops score while producing the worst breadth result in the recorded Qwen3-4B lineage. The public weights are useful as a reproducible warning: breadth gates must be frozen before specialist distillation.

distillation, tool calling, negative transfer, rejected

## Headline Numbers

Evidence class applies to each value: measured = retained evaluation or
runtime result; historical = recorded result not freshly reproduced;
derived = calculated or decided from records; observed = current artifact
state; not-measured = an explicit evidence gap.

**File-ops depth:** 100%

saturated narrow gate

Evidence: historical

**Breadth:** 31%

down from 59.6% stock

Evidence: historical

**Breadth delta:** -28.6pp

recorded stock-to-candidate change

Evidence: derived

**Decision:** Reject

failed negative-transfer artifact

Evidence: derived

## Competitive Context

| System | Metric | Score | Size / Class | Comparable? | Readout |
| --- | --- | --- | --- | --- | --- |
| Multibackend-distilled 4B | recorded depth / breadth | 100% / 31% | 4B | Direct | The target gate stayed saturated while breadth collapsed. |
| Stock Qwen3-4B | same recorded depth / breadth family | 58% / 59.6% | 4B | Direct | The candidate gained depth but lost 28.6 breadth points. |
| File-ops-distilled 4B | same recorded depth / breadth family | 100% / 42.3% | 4B | Direct | Already showed negative transfer; multibackend distillation made breadth worse. |
| ReST-fused 4B | same recorded depth / breadth family | 100% / 65% | 4B | Direct | The later teacher-free ReST attempt is the breadth-preserving research candidate. |

Direct rows share this artifact's eval setup. Directional rows are
useful market context but should not be read as leaderboard claims.

## Attempts and decisions

| Attempt | Method | Depth | Breadth | Decision |
| --- | --- | --- | --- | --- |
| A0 | Stock Qwen3-4B | 58% | 59.6% | Baseline |
| A1 | File-ops distillation | 100% | 42.3% | Route only |
| A2 | Multibackend distillation | 100% | 31% | Reject |
| A3 | Teacher-free ReST | 100% | 65% | Research-only routed ship |

## Release Blockers

### Breadth collapsed

The recorded 31% breadth result is 28.6 points below stock and 11.3 points below the narrower file-ops distillation.

**Unblock:** Do not promote this checkpoint. Start from a breadth-preserving recipe with a frozen regression gate.

### No committed machine-readable eval package

The exact historical result is preserved in the attempt ledger and session record, but raw predictions and a package-level eval report were not committed.

**Unblock:** Retain the result as historical evidence; only rerun if a new recipe needs the same checkpoint as a controlled baseline.

## Evidence

- [Hugging Face model](https://huggingface.co/posttrainllm/qwen3-4b-multibackend-distilled)

- [Attempt ledger](https://posttrainllm.com/docs/attempt-ledger)

- [Historical model inventory](https://posttrainllm.com/docs/sessions/06-17-stepback-inventory-roi/)

- [Public artifact registry](https://posttrainllm.com/docs/factory/public-artifacts)

## Next Release Action

Keep the weights public as a failed comparison artifact. Do not spend compute revalidating them unless a new breadth-preserving recipe explicitly needs this checkpoint as its baseline.
