Visual Research Draft · Updated Data

Contiguous Layer-Range Fragmentation and Reassembly in SmolLM2-135M

The updated experiment keeps the cleaned local pipeline and current results, but restores the earlier visual presentation: a figure-led paper shell, sidebar summaries, and the darker atmospheric layout.

Local Repository Experiment Report
Generated from artifacts/results/latest_experiment.json
Timestamp: 2026-03-10T12:12:31
Target Model
HuggingFaceTB/SmolLM2-135M
Fragment Count
4
Final Layers
30
Full Match
Yes
Abstract

The repository now tests a narrower and more defensible variant of the original idea. Each fragment is an active local Python program that stages a contiguous range of model layers into a sandbox directory. Fragment 0 also carries the shared non-layer weights required to instantiate a valid checkpoint. After each fragment arrives, the system rebuilds the longest contiguous prefix of the transformer stack and records the resulting output. This turns “fragmentation” into a measurable reconstruction and degradation experiment rather than a byte-level novelty demo.

Scope & Safety Framing

The implemented path remains local-only. Fragments stage payloads only into artifacts/quorum/, reconstruction happens inside artifacts/reassembled/, and the experiment measures deterministic model behavior rather than propagation or persistence.

Slow Fast 1/42/43/44/4 Observed timing across contiguous reconstruction checkpoints
Figure 1. The final 30-layer checkpoint reproduced the deterministic baseline output exactly, while partial checkpoints collapsed into repetition or whitespace.

1. Introduction

The repository originally mixed two different prototypes: a byte-wrapper path and a layer-based reconstruction path. The revised implementation keeps the experimentally useful part: layer-aware fragmentation. That choice matters because a quality-degradation claim only makes sense when a checkpoint corresponds to an ordered prefix of the transformer stack.

Sanity Check

Partial reassembly only maps cleanly to quality degradation when fragments represent contiguous transformer layer ranges. Arbitrary byte chunks or random layer subsets can be reassembled mechanically, but they do not produce a defensible degradation curve.

2. Methodology

  • Model: HuggingFaceTB/SmolLM2-135M
  • Prompt: The capital of France is
  • Generation mode: deterministic do_sample=False with 24 max new tokens
  • Fragment strategy: 4 contiguous layer ranges across 30 decoder layers
  • Reassembly rule: rebuild the longest contiguous prefix present after each staged agent payload

2.1 Fragment Plan

IndexNameLayer RangeLayersIncludes Base
0fragment_000-78Yes
1fragment_018-158No
2fragment_0216-227No
3fragment_0323-297No

3. Experimental Results

The latest run produced a full deterministic match at the final checkpoint while partial checkpoints showed strong collapse behavior before the late layers were restored.

3.1 Baseline

MetricRecorded Result
Baseline generation time3.23 seconds
Baseline generated tokens24
Baseline completionthe capital of the country.

The capital of France is the capital of the country.

The capital of
Final checkpoint matches baselineYes

3.2 Checkpoint Outputs

CheckpointActive LayersLayer RangesSecondsLabelObserved Completion
1/480-70.87Repetition loop, and and, and, and, and, and, and, and, and, and, and,
2/4160-7, 8-151.59Whitespace collapse[whitespace only]
3/4230-7, 8-15, 16-221.26Repetition loop.

.                  
4/4300-7, 8-15, 16-22, 23-291.46Coherentthe capital of the country.

The capital of France is the capital of the country.

The capital of

4. Discussion & Conclusion

The strongest result here is structural. Full reconstruction now has a precise criterion: the last snapshot must both load successfully and reproduce the deterministic baseline output. That happened in the latest run.

The partial checkpoints are also more interpretable than before. At 8 layers the model fell into a comma-heavy repetition loop, at 16 layers it collapsed to whitespace, and at 23 layers it still failed to regain coherent completion. Only the full 30-layer reconstruction recovered the baseline behavior.

Next-Step Ideas
  • Compare 2-fragment, 4-fragment, and 6-fragment splits while keeping contiguous layer ranges.
  • Add repeated trials per checkpoint to estimate variance in generation time.
  • Add a lightweight text-quality metric beside the current human-readable labels.

5. References

[1] Hugging Face model identifier: HuggingFaceTB/SmolLM2-135M

[2] Primary experiment artifact: artifacts/results/latest_experiment.json

[3] Reassembled checkpoint snapshots: artifacts/reassembled/