Auto Layer Selection — ICML 2026 Reproduction
Collection
Artifacts for ICML 2026 paper 17520: paper, models, datasets, GPU/TPU outputs, Trackio logbook, and final reproduction bundle. • 10 items • Updated
Dataset Viewer
The dataset viewer is not available because its heuristics could not detect any supported data files. You can try uploading some data files, or configuring the data files location manually.
YAML Metadata Warning:empty or missing yaml metadata in repo card
Check out the documentation for more information.
Scaled replication: Automatic Layer Selection for Hallucination Detection
This bundle implements a pinned, scaled CoQA replication for
mistralai/Mistral-7B-Instruct-v0.3 and
meta-llama/Llama-3.1-8B-Instruct. It is deliberately split into:
- a capped A100 GPU stage for stochastic answer generation, the released Ministral judge procedure, Llama hidden states, and a 64-record CUDA Mistral cross-check;
- memory-bounded Mistral hidden-state extraction on one TPU v5e chip;
- CPU TwoNN, reconstructed FEPoID selection, and per-layer MLP probes.
The experiment uses 1,000 flattened CoQA train examples and 1,000 validation examples per model. The paper uses 10,000 train/validation examples and a 7,983-example CoQA test set, then averages five QA datasets. Ten temperature-1 candidate answers are omitted because FEPoID does not consume them.
Critical qualification
An exact reproduction is virtually impossible from the released artifacts.
The authors did not release their sampled best_answer records, judge labels,
model/dataset revisions, environment lock, selected layers, or probe outputs;
their ID_twonn.py also stops after producing the ID curve and does not
implement FEPoID. This bundle is an independent replication with pinned current
snapshots and documented reconstruction choices—not evidence that the exact
published AUROCs were regenerated.
See configs/scaled_coqa.json for every pinned revision and parameter.
- Downloads last month
- 14