Datasets:
id stringlengths 36 36 | tier stringclasses 1
value | github_login stringclasses 179
values | created_at stringlengths 32 32 | status stringclasses 2
values | sha256 stringlengths 64 64 | source_bytes int64 341 260k | leaderboard_rank int64 | source large_stringlengths 341 260k | metadata_json large_stringlengths 645 734 | result_json large_stringlengths 5 4.8k |
|---|---|---|---|---|---|---|---|---|---|---|
0001d8fa-3147-4fd0-a49d-61614c52d9ac | easy | CodeReclaimers | 2026-08-20 15:21:07.916546+00:00 | succeeded | ecdc5bc3549f21ec4b5703f65802288e92bc05cd2c1587f9346ab4e680aba37d | 42,274 | null | """v39_keyed: keyed-modulus CRT learner — Run A of T1_CERT_PLAN_20260819.
HYPOTHESIS A (alias modulus): h1's displayed N is a stable per-identity
alias phi(M) of a hidden modulus M <= 2^20 from enumerated factor pairs
(<=10-bit primes), while x and y are displayed plainly and the step closes
per prime channel (squarin... | {
"id": "0001d8fa-3147-4fd0-a49d-61614c52d9ac",
"created_at": "2026-08-20 15:21:07.916546+00:00",
"db_md5": "4734e439309d100106007d6b654ccab0",
"submitter": "CodeReclaimers",
"github_login": "CodeReclaimers",
"run_id": "9824eccf-75ad-45fa-86fd-90000b08b1dc",
"tier": "easy",
"dataset_id": "e5",
"status... | {
"score": {
"mean_loss": 1.6747209675214556,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.47916666706403094
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.2111226555477876,
"example_count": 600... |
000228af-ba9c-4313-a87f-74121f716157 | easy | liam-gb | 2026-08-11 10:17:03.181174+00:00 | succeeded | a52fd23b420d06c1ea43204544d5e9ffbcb40ea72e22fa95b26f8d1eaa5e17e6 | 12,304 | null | """rns_critical_uni: critical-path loop with uniform per-loop supervision.
Each field value enters as its residue phase on a bank of small odd prime
circles via fixed sinusoidal feature maps over digit significance
(omega[s] = 2*pi*(10^s mod p)/p) — carry-free by construction. Phases are
resolved against fixed unit-ci... | {
"id": "000228af-ba9c-4313-a87f-74121f716157",
"created_at": "2026-08-11 10:17:03.181174+00:00",
"db_md5": "4e4e60e58520faa9b1efeba2156ec5ae",
"submitter": "liam-gb",
"github_login": "liam-gb",
"run_id": "15476376-a505-4f85-ac7e-e1c77ec3e5bb",
"tier": "easy",
"dataset_id": "e1",
"status": "succeeded"... | {
"score": {
"mean_loss": 5.046521425247192,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0800000000745058
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 7.268509864807129,
"example_count": 100,
... |
0009d371-ce6b-4cb1-8742-acab0548134c | easy | jordanrubin | 2026-08-06 19:23:31.320469+00:00 | succeeded | 2214473aabeee3e5a3252928aedd50993436405bf900fc579ecc4ef88811f4f4 | 14,788 | null | """Autonomous digit-state recurrent Transformer for repeated modular squaring.
The model is deliberately organized around one learned transition:
p_0 = right_aligned_decimal_digits(x)
p_{k+1} = F_theta(p_k, decimal_digits(N))
The same two-block Transformer cell is applied exactly T times. T is used only
as ... | {
"id": "0009d371-ce6b-4cb1-8742-acab0548134c",
"created_at": "2026-08-06 19:23:31.320469+00:00",
"db_md5": "f917ba4387feee357d1f6822e4a7825f",
"submitter": "Jordan Rubin",
"github_login": "jordanrubin",
"run_id": "7a653879-d845-4735-89f6-8d24230933d3",
"tier": "easy",
"dataset_id": "e3",
"status": "s... | {
"score": {
"mean_loss": 2.1296750745907396,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0018749999580904841
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.1309655856123277,
"example_count": 8... |
000a3457-d237-4459-bed2-b08371ee3d12 | easy | sapient-sapiens | 2026-08-19 19:24:48.910974+00:00 | succeeded | daa4fc1f1821b95b049e62d1602f10121592d66b9efbbb3e78e141bca4403b2b | 33,252 | null | """Depth-quantized recurrence with a mixture-over-depth objective.
The recurrent state is a bank of right-aligned digit slots. Every operator
application is followed by a soft quantization back onto the token simplex, and
the digit logits produced by that quantization are the answer logits at that
depth. There is no s... | {
"id": "000a3457-d237-4459-bed2-b08371ee3d12",
"created_at": "2026-08-19 19:24:48.910974+00:00",
"db_md5": "2553f964fd1803486430c2690edf5c49",
"submitter": "Ertondy",
"github_login": "sapient-sapiens",
"run_id": "bc701422-a4a4-4e0b-a61a-aa3779c48292",
"tier": "easy",
"dataset_id": "e3",
"status": "su... | {
"score": {
"mean_loss": 2.2663152426947795,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.008125
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.276738232607394,
"example_count": 800,
... |
000c7da0-a4e9-4f0d-ac6d-ceba3bbd79e3 | easy | karanganesan | 2026-08-06 05:45:58.399132+00:00 | succeeded | 35bb515be06389766e7968d9741d65099fd39cb897f018b8a7b9b4d9e5188a5a | 25,616 | null | """Parametric looped-transformer family (P1).
One weight-tied transformer block applied k times in latent space. Config
flags cover four P1 families with one file:
looped recall=0 gated=0 tfilm=0 plain weight-tied loop
looped-recall recall=1 re-inject the input embedding each
... | {
"id": "000c7da0-a4e9-4f0d-ac6d-ceba3bbd79e3",
"created_at": "2026-08-06 05:45:58.399132+00:00",
"db_md5": "853fe7df67aa06b7473fbcdd5ee2d312",
"submitter": "Karan Ganesan",
"github_login": "karanganesan",
"run_id": "9ce87288-351f-43fc-9286-32c267afb2c7",
"tier": "easy",
"dataset_id": "e1",
"status": ... | {
"score": {
"mean_loss": 3.7291706800460815,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.011666666716337204
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 3.6878952980041504,
"example_count": 10... |
0019ad4c-eb58-4195-91c7-8648d5bcbcb6 | easy | ArkinDharawat | 2026-08-15 20:08:59.860054+00:00 | succeeded | b3a3fc7ccd6a67634a1aaadb13d50e511a3f9ba42651a0b7a87cef3c72bf10ef | 7,633 | null | """Universal-Transformer-style looped block, v2: step embed + pos offset + LR schedule + 6 loops.
Extends ``looped_ut_padded`` with four ideas drawn from
Graves 2016 (ACT), Dehghani et al. 2018 (Universal Transformer), and
Merrill & Sabharwal 2025 (log-depth transformers):
1. **Learned per-step (timestep) embedding**... | {
"id": "0019ad4c-eb58-4195-91c7-8648d5bcbcb6",
"created_at": "2026-08-15 20:08:59.860054+00:00",
"db_md5": "df88b48f30c8035a9269455deffdd66b",
"submitter": "Arkin Dharawat",
"github_login": "ArkinDharawat",
"run_id": "d7867b75-a51b-4a0b-8db9-6c11320331f7",
"tier": "easy",
"dataset_id": "e3",
"status"... | {
"score": {
"mean_loss": 2.240777682338874,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.009375000055879355
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.2808395810221542,
"example_count": 800... |
001d49d5-358e-4e45-9cab-9a698d534eb3 | easy | mnida | 2026-08-25 22:18:55.230889+00:00 | succeeded | a18891d3780d413e19103a57d2fe22a3de5a430300e068f22fb0df54e9f2b708 | 15,662 | null | "\"\"\"Large quadratic looped Transformer with immutable direct-N-blind routes.\"\"\"\n\nfrom __futu(...TRUNCATED) | "{\n \"id\": \"001d49d5-358e-4e45-9cab-9a698d534eb3\",\n \"created_at\": \"2026-08-25 22:18:55.230(...TRUNCATED) | "{\n \"score\": {\n \"mean_loss\": 2.1489941186962733,\n \"primary_metric\": \"mean_exact_acc(...TRUNCATED) |
001e4315-d76b-43da-8a88-0f8a9d2e3f95 | easy | nikolageorgiev2000 | 2026-08-04 15:19:10.646519+00:00 | succeeded | bcb781bfdf7ba38a0a37eaa6ed064a66f4dfd339aab3b052a335c7c7073403a1 | 26,498 | null | "\"\"\"T-composed categorical-digit reasoner with fully learned pair/reduction maps.\"\"\"\n\nfrom _(...TRUNCATED) | "{\n \"id\": \"001e4315-d76b-43da-8a88-0f8a9d2e3f95\",\n \"created_at\": \"2026-08-04 15:19:10.646(...TRUNCATED) | "{\n \"score\": {\n \"mean_loss\": 2.206954932994406,\n \"primary_metric\": \"mean_exact_accu(...TRUNCATED) |
00224bd0-73fb-4a40-97f8-a9b07f9da26e | easy | shreyash-chonkie | 2026-08-26 14:24:03.669929+00:00 | succeeded | c55d3c9a0c2334b02bb81847ede9332c2e7c70fa8081fb22b066ce23f236b91d | 11,398 | null | "\"\"\"Ternary register machine with supervised execution prefixes.\"\"\"\n\nfrom __future__ import (...TRUNCATED) | "{\n \"id\": \"00224bd0-73fb-4a40-97f8-a9b07f9da26e\",\n \"created_at\": \"2026-08-26 14:24:03.669(...TRUNCATED) | "{\n \"score\": {\n \"mean_loss\": 1.944562554359436,\n \"primary_metric\": \"mean_exact_accu(...TRUNCATED) |
0029af78-e042-41d6-8ff6-f63b473acf21 | easy | erdavis0 | 2026-08-22 07:43:23.646040+00:00 | succeeded | a1c22e3f3f7475c81762f81ac0379de4cfc7d9c3c89e6889339eb862577f0cc4 | 6,485 | null | "\"\"\"Field crossbar with learned row and column summaries.\"\"\"\n\nfrom __future__ import annotat(...TRUNCATED) | "{\n \"id\": \"0029af78-e042-41d6-8ff6-f63b473acf21\",\n \"created_at\": \"2026-08-22 07:43:23.646(...TRUNCATED) | "{\n \"score\": {\n \"mean_loss\": 6.961434841156006,\n \"primary_metric\": \"mean_exact_accu(...TRUNCATED) |
One Layer Deeper submissions
This dataset archives 15,602 accepted uploads from 206 GitHub accounts to the One Layer Deeper competition. It contains 9,627 distinct source files, all upload metadata, and stored evaluation results. Snapshot: September 7, 2026, 21:48 UTC, after the August 31 submission deadline.
| Split | Uploads | Succeeded | Failed |
|---|---|---|---|
| easy | 11,961 | 11,112 | 849 |
| medium | 2,704 | 2,509 | 195 |
| hard | 937 | 847 | 90 |
All accepted uploads are included: practice runs, failures, identical resubmissions, and superseded or excluded entries. “Succeeded” is the evaluator's run status, not a claim of task correctness or rule compliance. GitHub accounts do not necessarily represent distinct people. Rejected requests and local experiments are outside this archive.
Load the data
This is a private dataset under GPUMODE. Access requires an authorized Hugging Face account. Run hf auth login first; datasets uses that local authentication for the download.
from datasets import load_dataset
hard = load_dataset("GPUMODE/one-layer-deeper-submissions", split="hard")
row = hard[0]
print(row["id"], row["github_login"], row["status"])
print(row["source"][:500])
Install the datasets package to use this example. Loading reads participant source as text; executing that source is unnecessary for archive analysis.
Each row has these fields:
| Field | Meaning |
|---|---|
id |
Accepted submission UUID |
tier |
easy, medium, or hard |
github_login |
Submitting account label |
created_at |
Stored upload timestamp, including UTC offset |
status |
Stored evaluator run status |
sha256 |
SHA-256 of the original UTF-8 source bytes |
source_bytes |
Original source length in bytes |
leaderboard_rank |
Rank of this exact upload in the archived public leaderboard, or null |
source |
Complete original submission.py decoded as UTF-8, preserving line endings |
metadata_json |
Complete original metadata.json file as a UTF-8 string, preserving formatting |
result_json |
Complete original result.json file as a UTF-8 string, preserving formatting |
A failed run may have JSON null in result_json; use json.loads(row["result_json"]) to interpret it. Scores are recorded evaluator outcomes, not independent reproductions. A high score does not establish rule compliance or generalization beyond the recorded tests.
Original files and integrity
data/{easy,medium,hard}.parquet One row per accepted upload
archives/{easy,medium,hard}.tar.gz
submissions/<tier>/<id>/submission.py
submissions/<tier>/<id>/metadata.json
submissions/<tier>/<id>/result.json
inventory/manifest.json
inventory/export_summary.json
inventory/statistics.json
inventory/engagement.json
inventory/leaderboard.json
verification.json Local packaging verification report
checksums.json File SHA-256 hashes and sizes
SHA256SUMS Checksums of every other packaged file
The tar archives preserve all three files byte for byte. Archive order and timestamps are normalized; the files' content bytes are unchanged. Parquet strings also round-trip exactly to the original bytes via UTF-8 encoding, including CRLF line endings and JSON whitespace.
import hashlib
import json
source_bytes = row["source"].encode("utf-8")
assert len(source_bytes) == row["source_bytes"]
assert hashlib.sha256(source_bytes).hexdigest() == row["sha256"]
metadata = json.loads(row["metadata_json"])
result = json.loads(row["result_json"])
Packaging verified every Parquet source hash, every metadata/result string against its original bytes, every archive entry against its original file, and exact submission-ID coverage in each split. checksums.json maps every substantive file path to its SHA-256 and byte length, excluding the two checksum manifests. SHA256SUMS additionally covers checksums.json; it does not hash itself. With the repository downloaded locally, run sha256sum -c SHA256SUMS (or shasum -a 256 -c SHA256SUMS on macOS).
Provenance and scope
The original export used a read-only, repeatable-read transaction against the organizer database. Every source matched the database's md5(source) value, and SHA-256 was saved per upload. The public leaderboard was fetched separately immediately before that transaction. inventory/export_summary.json records the snapshot and exclusions; inventory/manifest.json retains per-upload provenance.
This release explicitly includes only the submission files, five listed inventory files, and packaging documentation. It does not include dataset inputs, trained checkpoints, raw service logs, metric histories, moderation records, credentials, or email fields. Participant source is retained as submitted. Moderation status is not a row-level classification in this release, and absence from the leaderboard alone does not establish a reason for exclusion.
Participant files retain their existing ownership and terms. No blanket license or relicensing is asserted for these uploads. The upstream service's Apache license does not automatically apply to participant submissions.
The packaging script reads code as bytes and text and never imports or executes participant source. Rebuilding with the same inputs and Python/PyArrow versions produces deterministic archive content and packaging metadata; the exact runtime versions are recorded in verification.json.
- Downloads last month
- 7