forgetting
self-training-denoising-forgetting
Self-Training: Denoising vs. Forgetting — Reproduction Artifacts
Reproduction artifacts for "Why Self-Training Helps and Hurts" (arXiv:2602.14029):
the U-shaped risk curve from the denoising–forgetting trade-off, in overparameterized
linear regression (synthetic, spiked covariance) and a small CIFAR-10 deep-learning analogue.
Read REPORT.md first — it states which claims reproduced, with what numbers,
and lists all assumptions, deviations, hardware and cost.
Contents… See the full description on the dataset page: https://huggingface.co/datasets/pngwn/self-training-denoising-forgetting.vlm-forgetting-datasetsmbpp-code-rl
MBPP for code RL (deduplicated against MBPP+)
MBPP prepared for RLVR training in verl,
with two independent hold-outs so both MBPP+ and MBPP's own canonical test
split stay reportable after training on this data.
split
rows
contents
train
320
MBPP canonical train + validation + prompt, minus everything in MBPP+
test
378
exactly the problems in evalplus/mbppplus
heldout_mbpp_test
276
MBPP's canonical test split (task_id 11-510) that is not in MBPP+… See the full description on the dataset page: https://huggingface.co/datasets/RL-Forgetting-Experiments-3/mbpp-code-rl.synthetic-selftrain-denoising-forgetting
Synthetic self-training denoising-vs-forgetting trajectories
Reproduction of the U-shaped risk curve from Wu, Yang & Sun,
Why Self-Training Helps and Hurts: Denoising vs. Signal Forgetting
(arXiv:2602.14029; ICML 2026 submission,
OpenReview VnA5q5jXVz).
This dataset contains all configs, per-iteration trajectories, theoretical
predictions, and empirical risk measurements for the synthetic
overparameterized-linear-regression self-training experiments. Every point on
every curve… See the full description on the dataset page: https://huggingface.co/datasets/evalstate/synthetic-selftrain-denoising-forgetting.denoising-forgetting-linear-repro2025-rethinkdc-imagenet-forgetting-ipc-100
Dataset used for paper -> "Rethinking Dataset Compression: Shifting Focus From Labels to Images"
Dataset created according to the paper An Empirical Study of Example Forgetting during Deep Neural Network Learning.
Basic Usage
from datasets import load_dataset
dataset = load_dataset("he-yang/2025-rethinkdc-imagenet-forgetting-ipc-100")
For more information, please refer to the Rethinking-Dataset-Compression