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 |
|---|---|---|---|---|---|---|---|---|---|---|
3b73865f-4b31-42d7-a4b8-0c153a969714 | easy | newjordan | 2026-08-09 14:53:27.904386+00:00 | succeeded | 42bc8ea2737744d506f0c08a6128c112ff628e5a6b682f0a8ede8a96b63a1eca | 93,586 | null | """Helix-Crawler — a recursive-weight-shared "crawler" ALU for modular squaring.
The task is r_{i+1} = r_i^2 mod n, chained T times: a *recurrence* over a single
operator. This build embodies the crawler / Frugendorff lineage (spark:~/sota_crawler)
mapped onto that structure. n never enters as a table row (unseen modu... | {
"id": "3b73865f-4b31-42d7-a4b8-0c153a969714",
"created_at": "2026-08-09 14:53:27.904386+00:00",
"db_md5": "9533d0ab11c33c0209f25683f45cd086",
"submitter": "Frosty40",
"github_login": "newjordan",
"run_id": "eea58325-610a-4dd8-a7cb-3c4b6ce20aa5",
"tier": "easy",
"dataset_id": "e1",
"status": "succeed... | {
"score": {
"mean_loss": 0.03638477437198162,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 1.0
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 0.03660258278250694,
"example_count": 100,
"e... |
3b7e7cbb-848c-4120-b6ae-813240b8192e | easy | Dandandan | 2026-08-04 15:05:56.945774+00:00 | failed | 963c670cbc4f31753a8d47ba297ad442b876b570c15168cbe42c39eec09b2876 | 18,565 | null | """Multi-start learned spectral recurrence for One Layer Deeper.
Every coefficient that determines a prediction is randomly initialized and
updated end to end. Shared candidate transitions are unrolled to the prompt's
training depth; self-calibration and an integer-boundary loss favor a reusable
functional root withou... | {
"id": "3b7e7cbb-848c-4120-b6ae-813240b8192e",
"created_at": "2026-08-04 15:05:56.945774+00:00",
"db_md5": "6258bf441c11598637da3c01c37a28d1",
"submitter": "Daniël Heres",
"github_login": "Dandandan",
"run_id": "f3834fc5-76b3-46f4-b004-24351ed41434",
"tier": "easy",
"dataset_id": "e1",
"status": "fai... | null
|
3b8b1eca-de2d-495f-9a34-ffac0abda7e9 | easy | DDanlov | 2026-08-21 20:23:25.410721+00:00 | succeeded | e2de13bb1739fc015fcbfa412eaaa9018bd333b2db7da6968f09b14228a5021e | 26,066 | null | from __future__ import annotations
import math
import time
from typing import Any, Optional, Tuple, Dict
import torch
import torch.nn as nn
import torch.nn.functional as F
from torch.optim import Optimizer
try:
from benchmark.api import ModelSpec, OptimizerBundle, OptimizerSpec, Submission, assert_model_state
exce... | {
"id": "3b8b1eca-de2d-495f-9a34-ffac0abda7e9",
"created_at": "2026-08-21 20:23:25.410721+00:00",
"db_md5": "a84b843f8386d7529aa6c1c08dd7d6e9",
"submitter": "DDanlov",
"github_login": "DDanlov",
"run_id": "d8441c51-1d86-484a-b3c0-048a81de3712",
"tier": "easy",
"dataset_id": "e5",
"status": "succeeded"... | {
"score": {
"mean_loss": 2.819373554821347,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.00375
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.8203729639673445,
"example_count": 600,
"... |
3b8fee4d-6bac-4d4a-98f2-abf36ffaa0a6 | easy | DDanlov | 2026-08-28 21:46:47.667251+00:00 | succeeded | 6e04f2896b22fe52315c23bf0017b4fc32f7da5aee44411d458fe534d8b79932 | 14,935 | null | """
Trial 16: Model B: D=256, H=4, M=8, u=4 (T=32)
"""
import math
import torch
import torch.nn as nn
import torch.nn.functional as F
from dataclasses import dataclass
from typing import Optional, Tuple, Dict, Any, List, Union
try:
from benchmark.api import ModelSpec, OptimizerBundle, OptimizerSpec, Submission, T... | {
"id": "3b8fee4d-6bac-4d4a-98f2-abf36ffaa0a6",
"created_at": "2026-08-28 21:46:47.667251+00:00",
"db_md5": "de5c9b703f264f36871b372e38ffdf4b",
"submitter": "DDanlov",
"github_login": "DDanlov",
"run_id": "09839ddf-71f9-4e63-af41-801670d9e3a6",
"tier": "easy",
"dataset_id": "e5",
"status": "succeeded"... | {
"score": {
"mean_loss": 2.83321353302519,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0008333333333333334
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.8332135109356167,
"example_count": 600... |
3b8ffff4-e050-48e8-8326-485d4da4847b | easy | khushidahi | 2026-08-20 03:45:33.796786+00:00 | succeeded | e8d0ef1293bf84b563edcfdaa05060443ab1ef7b952bccf669827aea955103e3 | 33,081 | null | """R296: factorized tied causal categorical functional root.
This candidate is deliberately diagnostic. It trains on evaluator-owned final
labels at several depths so that one shared transition must compose. The model
never converts N or X to native numbers and never implements multiplication or
modular reduction.
... | {
"id": "3b8ffff4-e050-48e8-8326-485d4da4847b",
"created_at": "2026-08-20 03:45:33.796786+00:00",
"db_md5": "2cc0834a43cb60913a23bbf99bfa10af",
"submitter": "khushidahi",
"github_login": "khushidahi",
"run_id": "80945666-3255-4aba-bca6-ed1bf3e0da71",
"tier": "easy",
"dataset_id": "e3",
"status": "succ... | {
"score": {
"mean_loss": 2.2465507787234236,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.007500000009313226
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.2230593322654406,
"example_count": 80... |
3b926369-f420-4873-bcef-90359803adc7 | easy | k-penchev | 2026-08-22 21:11:21.732596+00:00 | succeeded | 3ee12bcb425a5f4e263e88cbe4d769efa9edcf1f77da0af9146d6bdffa87de51 | 4,809 | null | """Two-loop recurrent model with attention and a gated local digit mixer."""
from __future__ import annotations
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
assert_model_state,
)
D_MODEL ... | {
"id": "3b926369-f420-4873-bcef-90359803adc7",
"created_at": "2026-08-22 21:11:21.732596+00:00",
"db_md5": "fbf02c0c1a0d01a2fcfc252cc139f80d",
"submitter": "Kaloyan Penchev",
"github_login": "k-penchev",
"run_id": "b110bf47-5ac7-43a8-8b0d-7550bc727caf",
"tier": "easy",
"dataset_id": "e1",
"status": "... | {
"score": {
"mean_loss": 1.983376145362854,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0033333334140479565
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.057493209838867,
"example_count": 100... |
3b958835-2efd-4d08-9397-24e0da0878ef | easy | AdityaVG13 | 2026-08-15 10:23:40.620535+00:00 | succeeded | 00fe0f5c25cea41c426a1b265b06bdba47c9c06204d950b7d6a77b0a3062e214 | 110,011 | null | """AEGIS V11 v11_timeless: hopself JOIN with a time-invariant sequential leaf.
One Layer Deeper submission candidate. All task transitions remain learned
end-to-end. The deterministic shell only identifies marker-delimited roles,
constructs a finite semantic tape, schedules learned interval programs, and
applies a tri... | {
"id": "3b958835-2efd-4d08-9397-24e0da0878ef",
"created_at": "2026-08-15 10:23:40.620535+00:00",
"db_md5": "ca526f64196d118026d413999c16567b",
"submitter": "AdityaG",
"github_login": "AdityaVG13",
"run_id": "6eed317a-229f-4193-b39a-412b958ff1f1",
"tier": "easy",
"dataset_id": "e1",
"status": "succeed... | {
"score": {
"mean_loss": 1.9696411490440369,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.6500000059604645
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 1.1396230459213257,
"example_count": 100,... |
3ba0610f-39fd-4b9b-a94b-1595d1ffc84b | easy | yashkant | 2026-08-14 11:17:15.006844+00:00 | succeeded | 2ffbea7f8f884ee874af6956c4cab4f3b020a8ca574365b5b2b49dbba468d573 | 37,404 | null | """Round1863 learned state/T-conditioned BCH cubic coefficient.
Both arms keep the exact Round1851 AdamW 5e-3 / EMA .995 carrier and eager
BCH graph. A four-parameter bounded gate reads live correction geometry,
relative cubic size, and the public parsed T field at every recursive cycle.
The equal-work control compute... | {
"id": "3ba0610f-39fd-4b9b-a94b-1595d1ffc84b",
"created_at": "2026-08-14 11:17:15.006844+00:00",
"db_md5": "2b8256673c877c07bf3f5f413cf0def8",
"submitter": "Yash Kant",
"github_login": "yashkant",
"run_id": "8e156f87-94f2-4840-b8b9-199daf4f6b9a",
"tier": "easy",
"dataset_id": "e2",
"status": "succeed... | {
"score": {
"mean_loss": 2.7628480195999146,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.050833335146307945
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 3.2367613315582275,
"example_count": 30... |
3ba13abb-ce6d-4e86-8032-e3e5a62177cd | easy | karanganesan | 2026-08-08 04:03:27.809962+00:00 | succeeded | 8d9eecdc8607c91b190336e24b49597738decc69ce7d4e24ee5930ae97cfcad7 | 27,127 | 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": "3ba13abb-ce6d-4e86-8032-e3e5a62177cd",
"created_at": "2026-08-08 04:03:27.809962+00:00",
"db_md5": "3e8439ba8ab561733e7dc883fab4ecb2",
"submitter": "Karan Ganesan",
"github_login": "karanganesan",
"run_id": "a8d6dce5-91d7-40a8-b960-5ebfa49726a3",
"tier": "easy",
"dataset_id": "e4",
"status": ... | {
"score": {
"mean_loss": 7.184106826782227,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.005833333590999246
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 7.471048831939697,
"example_count": 1200... |
3bad690b-1db3-4304-8222-d2e34999a898 | easy | Yalyenea | 2026-08-18 18:55:27.845299+00:00 | succeeded | 7bf30e19d90e7f6cadb4ad67d0b2cd51634a4fd2d29a507be6f840a0dd78f8c1 | 3,628 | null | """Basic single-pass Transformer with PyTorch AdamW."""
from __future__ import annotations
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
assert_model_state,
)
D_MODEL = 128
NUM_HEADS = 4
... | {
"id": "3bad690b-1db3-4304-8222-d2e34999a898",
"created_at": "2026-08-18 18:55:27.845299+00:00",
"db_md5": "eaac69fe345f82973fcf392c504a40ff",
"submitter": "yfff",
"github_login": "Yalyenea",
"run_id": "0430eca5-7b70-4017-b8c4-5dc6fabd7623",
"tier": "easy",
"dataset_id": "e6",
"status": "succeeded",
... | {
"score": {
"mean_loss": 3.289332151412964,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.010135135613381863
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 4.37008810043335,
"example_count": 60,
... |
3bb26b00-7236-46ff-b196-a649bfe096c2 | easy | shreyash-chonkie | 2026-08-24 23:57:17.095605+00:00 | succeeded | d64bbafa4a82d5d78982aec75998bcfec3e3d416bb3f8c44ff63efcf77aa93f5 | 15,459 | null | """Recurrent attention with improvement-weighted pool feedback."""
from __future__ import annotations
import math
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
TokenLossBatch,
assert_mo... | {
"id": "3bb26b00-7236-46ff-b196-a649bfe096c2",
"created_at": "2026-08-24 23:57:17.095605+00:00",
"db_md5": "1253f45c068549baf3c3f4905f581fdf",
"submitter": "Shreyash",
"github_login": "shreyash-chonkie",
"run_id": "1d1467a0-5f32-447f-95b0-855d0fb8df1c",
"tier": "easy",
"dataset_id": "e8",
"status": "... | {
"score": {
"mean_loss": 2.8018022179603577,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.034090910106897354
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 3.715831756591797,
"example_count": 85,... |
3bb59fea-8f3e-4b09-b857-225736a56504 | easy | chad-atexpedient | 2026-08-27 01:12:51.322065+00:00 | succeeded | 7b428747d70a89070aedb70c7cff847d00e3e8401338e00b83c40ba3fa271ae0 | 17,905 | null | """X127: successor-aligned compositional state teacher.
X18 pairs a T=2 supplied output with an equal T=1 input, but compares both rows
after one transition, which are semantically offset. X127 computes one unmasked
shadow successor from each row's terminal state and trains the paired T=2 successor
toward the detached... | {
"id": "3bb59fea-8f3e-4b09-b857-225736a56504",
"created_at": "2026-08-27 01:12:51.322065+00:00",
"db_md5": "773a02788d4f6fb5a5b745faff32b9a2",
"submitter": "chad-atexpedient",
"github_login": "chad-atexpedient",
"run_id": "99be63ad-478d-4599-8f31-fdb11402e1cf",
"tier": "easy",
"dataset_id": "e5",
"st... | {
"score": {
"mean_loss": 2.279211227368614,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.013333333370586236
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.3265669281707217,
"example_count": 600... |
3bb5f5c0-835e-47ef-a2c1-49266b71bc4e | easy | ipsa-loquitur | 2026-08-26 04:37:43.849899+00:00 | succeeded | 204c59fd96a6ca9862d9343eebd71672cce61a33a50bc2f7827cc7c185709315 | 15,929 | null | """Standalone NetF-V1 submission candidate for One Layer Deeper.
The learned rule is shared across bit positions, modular-square iterations,
moduli, and tape widths. During training, exact double-and-add states provide
dense one-step supervision. During evaluation, the model receives no state
corrections and rolls out... | {
"id": "3bb5f5c0-835e-47ef-a2c1-49266b71bc4e",
"created_at": "2026-08-26 04:37:43.849899+00:00",
"db_md5": "7f53a54057b5ff2a1f13b4d4628c779a",
"submitter": "ipsa-loquitur",
"github_login": "ipsa-loquitur",
"run_id": "157dcb3a-875b-4574-82bb-d43f30d0c68c",
"tier": "easy",
"dataset_id": "e4",
"status":... | {
"score": {
"mean_loss": 5.906519109967574,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.7038541686534882
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 7.796711690163339,
"example_count": 1200,
... |
3bb77e77-bbb0-4ff3-b9f2-2cf4fc7ece21 | easy | Bananafly | 2026-08-12 09:17:28.485443+00:00 | succeeded | 5835a56ff941d251b1b59338285849b3701d3b6a5f1df96d7ce97ee52fb014d4 | 11,637 | null | """T-aligned recurrent soft decimal-residue model."""
from __future__ import annotations
import math
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
assert_model_state,
)
D_MODEL = 64
NUM_H... | {
"id": "3bb77e77-bbb0-4ff3-b9f2-2cf4fc7ece21",
"created_at": "2026-08-12 09:17:28.485443+00:00",
"db_md5": "9c5bda9e79705f2e75dbf579b1243231",
"submitter": "Andre Kreidemann",
"github_login": "Bananafly",
"run_id": "44b88bb1-e29e-4ad7-932c-1429500de460",
"tier": "easy",
"dataset_id": "e1",
"status": ... | {
"score": {
"mean_loss": 2.679015874862671,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.6759746074676514,
"example_count": 100,
"exac... |
3bb8f209-8f8e-4740-8cf6-8042f70fa78d | easy | KaustubhKumar05 | 2026-08-26 02:46:48.205161+00:00 | succeeded | c98fd0fd4b378e6715a0a7bb9645aefec8040e758f35c2ec9a6f0cdb19d6fb20 | 13,654 | null | """Neural-GPU-style conv GRU, v2s: v2 + speedrun pack.
- LR: warmup, hold max to 85% of wall-clock, steep late cliff (2026 schedule theory)
- identity-at-init: cell output projection zero-initialized
- batch reuse x6 (multi-pass theory)
Base (v2):
Changes over v1 (each backed by measured results in the literature):
... | {
"id": "3bb8f209-8f8e-4740-8cf6-8042f70fa78d",
"created_at": "2026-08-26 02:46:48.205161+00:00",
"db_md5": "68e1cb2c15c6308aca244e5732e3be95",
"submitter": "koz",
"github_login": "KaustubhKumar05",
"run_id": "3e51cc8a-1b27-40a1-8450-8c3e2215d8a1",
"tier": "easy",
"dataset_id": "e6",
"status": "succee... | {
"score": {
"mean_loss": 4.727572917938232,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.33693695813417435
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 7.1809492111206055,
"example_count": 60,
... |
3bbc01e0-36d9-44df-ba4a-f8d99a3246be | easy | lpbb | 2026-08-07 19:27:11.180913+00:00 | failed | 98556529fadd2e0c1d843680ab972281929828829bf682de4b26ea44edc4236d | 24,042 | null | """VARIANT R2b `fourier` — representation-prior axis (phase bank).
Where the campaign stands: memorization cannot be blocked (d=64 with 106k
params still memorizes 3200 rows in 45 s), delayed into irrelevance
(wd 1.0 at ~1000 full-batch epochs), or taxed away (softcap+smoothing
bounds confidence, map unchanged). The ... | {
"id": "3bbc01e0-36d9-44df-ba4a-f8d99a3246be",
"created_at": "2026-08-07 19:27:11.180913+00:00",
"db_md5": "957561b78131c4868751936af7afcbca",
"submitter": "Lpbb",
"github_login": "lpbb",
"run_id": "0b78b5ec-a2fc-41b3-a105-641a182dba50",
"tier": "easy",
"dataset_id": "e1",
"status": "failed",
"star... | null
|
3bbcea6f-4895-4faa-8c0d-c36a0ce58029 | easy | khushidahi | 2026-08-21 20:16:08.417026+00:00 | succeeded | 4d52f45d6cd81a0d78d06eaf117e0ec004717eed36edb878c9f8cea646636c24 | 29,148 | null | """R353: single-injection K8 variable-modulus control.
This submission tests the strongest rules-safe generic baseline suggested by
the public competition discussion. It does not implement squaring, modular
reduction, or any other task solver. Numeric prompt fields are only rearranged
onto equal-length, right-aligne... | {
"id": "3bbcea6f-4895-4faa-8c0d-c36a0ce58029",
"created_at": "2026-08-21 20:16:08.417026+00:00",
"db_md5": "92361a1268303c496912dc7406fa7d16",
"submitter": "khushidahi",
"github_login": "khushidahi",
"run_id": "4c2b7258-c181-4cfd-9293-88c3623c125d",
"tier": "easy",
"dataset_id": "e5",
"status": "succ... | {
"score": {
"mean_loss": 2.2620837845885573,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0037500000124176345
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.260877950309103,
"example_count": 60... |
3bbeba31-d239-440a-ab38-813f86404753 | easy | khushidahi | 2026-08-08 06:30:43.559417+00:00 | succeeded | 6a8b25be2ce902d42c25c2e35971c950d8c7df6bff2598ccab1069be8bde6e7d | 12,553 | null | """R45 full-budget, learned-symbol baseline.
This model intentionally avoids task-specific numerical decoding:
- no fixed digit values;
- no decimal-to-integer conversion;
- no FoNE parser;
- no solver-derived supervision.
The evaluator time allowance determines the LR schedule horizon. `max_steps`
is deliberately hi... | {
"id": "3bbeba31-d239-440a-ab38-813f86404753",
"created_at": "2026-08-08 06:30:43.559417+00:00",
"db_md5": "53006b37f26e02284a7ca3fbdccc09a3",
"submitter": "khushidahi",
"github_login": "khushidahi",
"run_id": "ca2cca2d-7f24-4cbb-bbe1-6a0cc0ee3e6d",
"tier": "easy",
"dataset_id": "e5",
"status": "succ... | {
"score": {
"mean_loss": 2.8869478798194512,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0066666666356225805
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.894322156906128,
"example_count": 60... |
3bbff0e9-73a3-4f95-9dd2-c2bfae00cd00 | easy | Cree0618 | 2026-08-29 06:36:05.749540+00:00 | succeeded | f7afa7f6572dcee1d5457e82f22509c0ef4ebb080818239922e651c795222d34 | 4,068 | null | """E007: E002 tied recurrence with an untied output head.
Single-variable ablation inspired by the TRM paper's finding that tying the
input embedding and output classifier can hurt small-data generalization.
Everything else matches E002: d_model=128, four applications of one shared
Transformer block, AdamW, default ev... | {
"id": "3bbff0e9-73a3-4f95-9dd2-c2bfae00cd00",
"created_at": "2026-08-29 06:36:05.749540+00:00",
"db_md5": "b46502ebdd4d8d59b0716a2f199fc1c6",
"submitter": "Cree0618",
"github_login": "Cree0618",
"run_id": "a4b5e7de-e106-4d4f-94d1-bc6eed8945f4",
"tier": "easy",
"dataset_id": "e5",
"status": "succeede... | {
"score": {
"mean_loss": 8.866152149227918,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.007083333370586236
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 9.047735021787908,
"example_count": 600,... |
3bc2c806-6ed3-4892-b97f-620b27162ea2 | easy | alirezashirvani-jr | 2026-08-30 21:55:57.175550+00:00 | succeeded | 996835699f04ca123f7f0f455750424f52a506091ca593b96781ae013da5dace | 24,944 | null | from __future__ import annotations
import math
import time
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import ModelSpec, OptimizerBundle, OptimizerSpec, Submission, TokenLossBatch, assert_model_state
EXPERT_COUNT = 12
STATE_COUNT = 64
PAIR_COUNT = EXPERT_COUNT * (EXPERT... | {
"id": "3bc2c806-6ed3-4892-b97f-620b27162ea2",
"created_at": "2026-08-30 21:55:57.175550+00:00",
"db_md5": "ae7bdcbfa8b36f64ed399730e2528184",
"submitter": "alirezashirvani-jr",
"github_login": "alirezashirvani-jr",
"run_id": "296da6f3-f768-48fb-ad67-71efe9086405",
"tier": "easy",
"dataset_id": "e3",
... | {
"score": {
"mean_loss": 5.237245559692383,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.008749999571591616
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 5.165926933288574,
"example_count": 800,... |
3bc4b272-23fd-403f-b43d-1b78e0ffb6f2 | easy | satyaki-up | 2026-08-07 17:02:22.411890+00:00 | succeeded | 4ef5a28d97249fcaebc993a2997307dfaae9b9787b689a629ed019fab7bdabdb | 11,493 | null | """Leaky LSTM with RNN optimization tricks from arXiv:1212.0901."""
from __future__ import annotations
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
assert_model_state,
)
WIDTH = 2151
LAYE... | {
"id": "3bc4b272-23fd-403f-b43d-1b78e0ffb6f2",
"created_at": "2026-08-07 17:02:22.411890+00:00",
"db_md5": "3c1ffdf55415da68af5af8b50a3dcff6",
"submitter": "Satyaki",
"github_login": "satyaki-up",
"run_id": "032b9e67-3443-4dba-9f80-a589b19ab556",
"tier": "easy",
"dataset_id": "e5",
"status": "succeed... | {
"score": {
"mean_loss": 2.262315259314434,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.007916666666666667
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.275359336570774,
"example_count": 600,... |
3bc7032a-f30c-4da7-b181-a4eb43ef9b21 | easy | jordanrubin | 2026-08-01 20:21:59.069778+00:00 | succeeded | 2c6adbc91db77091eaa1fdb53b2b15fdc6e05306608a2440be5e6a2aa2591b8a | 18,183 | null | """Looped bidirectional selective-SSM (Mamba-style), pure PyTorch.
The statefulness bet: input-dependent SSM dynamics along the token axis,
with the block weight-tied and applied NUM_LOOPS times in depth with input
injection — recurrence in depth is where the task's T-iteration lives.
The selective scan is computed i... | {
"id": "3bc7032a-f30c-4da7-b181-a4eb43ef9b21",
"created_at": "2026-08-01 20:21:59.069778+00:00",
"db_md5": "35fec302dbad6ab1de27b5092ede0b6a",
"submitter": "Jordan Rubin",
"github_login": "jordanrubin",
"run_id": "c4606533-ada8-4359-929a-5bf55d05f767",
"tier": "easy",
"dataset_id": "e5",
"status": "s... | {
"score": {
"mean_loss": 4.255239554390535,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.006666666517655054
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 4.345352022369881,
"example_count": 600,... |
3bd0a5bc-f07f-48e3-a761-e80b69bab065 | easy | khushidahi | 2026-08-15 14:45:41.978136+00:00 | succeeded | 564b92f90e7c4c74dea9d5c06b433f6b7343bd6707c9eec46a0f3323bd4d0375 | 16,527 | null | """R6 structured decimal-workspace model for One Layer Deeper.
The model does not implement multiplication, modular reduction, or the public
recurrence. It uses the public tokenizer structure to build learned, right-
aligned decimal tapes for N, X, and T, then applies a shared neural transition
to a mutable work tape.... | {
"id": "3bd0a5bc-f07f-48e3-a761-e80b69bab065",
"created_at": "2026-08-15 14:45:41.978136+00:00",
"db_md5": "72bedeec6e8753f631f179141ddc08bd",
"submitter": "khushidahi",
"github_login": "khushidahi",
"run_id": "1b4a2164-b058-4181-a905-e489336dc275",
"tier": "easy",
"dataset_id": "e5",
"status": "succ... | {
"score": {
"mean_loss": 2.188570499420166,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.00541666685603559
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.1874241828918457,
"example_count": 600,... |
3bd9a6bf-87f7-4fc0-ad5e-1a7d60c8991e | easy | rougee | 2026-08-22 06:36:19.583785+00:00 | failed | bc326e69f5345d564925eb98ff8d3d360dfc4f188505407418e31109b0cbfdb4 | 10,723 | null | """Official submittable bundle for One Layer Deeper.
Features:
- Dynamic prompt ingestion with robust tensor-based extraction of N, x, T, and ANS token positions.
- Sequential looped circuit depth recurrence supporting arbitrary depth (T in 1..64).
- Vectorized place-value digit decoding in pure PyTorch (exact multi-d... | {
"id": "3bd9a6bf-87f7-4fc0-ad5e-1a7d60c8991e",
"created_at": "2026-08-22 06:36:19.583785+00:00",
"db_md5": "492946d71bba8e182bfee2b25ef78857",
"submitter": "Rougee",
"github_login": "rougee",
"run_id": "22383e83-2280-4d1a-b1d0-a2c99160fb8f",
"tier": "easy",
"dataset_id": "e1",
"status": "failed",
"... | null
|
3be26006-89c1-48fe-845c-b6300ae75b96 | easy | blake-camp-surge | 2026-08-05 18:23:13.822506+00:00 | succeeded | 1d075d719b845e085f3f2aaaac92080c7e80038250b8914c83e953c60c9cc4c3 | 10,387 | null | """Vector-state neural cellular automaton over attention edges."""
from __future__ import annotations
import math
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
assert_model_state,
)
TOKEN... | {
"id": "3be26006-89c1-48fe-845c-b6300ae75b96",
"created_at": "2026-08-05 18:23:13.822506+00:00",
"db_md5": "b7ea6393bfe254e8f12aeb445fa260c6",
"submitter": "Blake Camp",
"github_login": "blake-camp-surge",
"run_id": "3bf6f821-05bd-4990-a39a-ae1503b6e5d0",
"tier": "easy",
"dataset_id": "e3",
"status":... | {
"score": {
"mean_loss": 3.40516807437367,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0025
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 3.516670226649572,
"example_count": 800,
"exa... |
3be7fc6e-645e-4117-9a06-440c4dbb5f5b | easy | oupadhyay | 2026-08-07 10:34:52.415717+00:00 | succeeded | 55d33f9eba92ed1d7b0516dd5b8b9f5d5cbef15996339f9224440e598758f263 | 7,984 | null | """Baseline Transformer plus diagnostic compressed-axial workspace creation."""
from __future__ import annotations
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
assert_model_state,
)
PAD, ... | {
"id": "3be7fc6e-645e-4117-9a06-440c4dbb5f5b",
"created_at": "2026-08-07 10:34:52.415717+00:00",
"db_md5": "a1f4bb439cb9e9ae6bcc9d56b1dc084f",
"submitter": "Ojasw Upadhyay",
"github_login": "oupadhyay",
"run_id": "9018e859-eb11-44ae-836f-ccf2c4329885",
"tier": "easy",
"dataset_id": "e5",
"status": "s... | {
"score": {
"mean_loss": 74.31122951336025,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0004166666666666667
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 75.17941961160155,
"example_count": 600... |
3bec19b2-ccd4-4b4b-94f9-87869bc17644 | easy | shirvani-jr | 2026-08-29 00:27:44.293769+00:00 | succeeded | 426a1ea4da9645aea9bb27937b6987cc1eaf0247c370e3cb0b2b652e9b3908b6 | 48,082 | null | """Adaptive learned discrete recurrent fabric for One Layer Deeper.
A gradient-trained, operation-free recurrent computational machine. The
model learns, from the competition's endpoint supervision: how to parse the
prompt into a bank of categorical digit registers, a reusable transition
circuit applied serially, a l... | {
"id": "3bec19b2-ccd4-4b4b-94f9-87869bc17644",
"created_at": "2026-08-29 00:27:44.293769+00:00",
"db_md5": "565694981a7f88014e65eadbd6483133",
"submitter": "Ali",
"github_login": "shirvani-jr",
"run_id": "c9403f99-8eac-49ec-9bab-b19700292fbc",
"tier": "easy",
"dataset_id": "e1",
"status": "succeeded"... | {
"score": {
"mean_loss": 1.872481882572174,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.07999999821186066
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 1.7722365856170654,
"example_count": 100,... |
3bed82a4-ea0c-4a22-ac37-99ff1a1afca7 | easy | adambrzosko | 2026-08-30 17:10:40.281164+00:00 | succeeded | aaad55260ceb883697d73364a0ae97ce4701670d33ce7e6775221922b98995a7 | 14,130 | null | """One Layer Deeper — digit-slot recurrent machine ("dm").
The prompt is [N] digits(N) [X] digits(x) [T] digits(T) and the label is the
decimal expansion of x^(2^T) mod N, tail-aligned: the last prompt position
carries the units digit, the one before it the tens digit, and so on.
Architecture. The model parses the ... | {
"id": "3bed82a4-ea0c-4a22-ac37-99ff1a1afca7",
"created_at": "2026-08-30 17:10:40.281164+00:00",
"db_md5": "fb033016a1b6cb8bea5f08ed98f913d8",
"submitter": "Adam Brzosko",
"github_login": "adambrzosko",
"run_id": "ee1d8197-5ce7-46ba-80c5-1e59a3f3d2f4",
"tier": "easy",
"dataset_id": "e5",
"status": "s... | {
"score": {
"mean_loss": 4.5336524472498425,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0104166666790843
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 4.494645170032176,
"example_count": 600,
... |
3bfc2977-d62e-44d7-8d16-c42b2bb5504b | easy | jordanrubin | 2026-08-06 19:50:08.081438+00:00 | succeeded | 09586f90ffa296e3bdbb25e6a0b059c7579344900d2f7d9eb12b0e902c4848f6 | 15,390 | null | """Bidirectional transformer, D=256 x 3 blocks, grokking-tuned AdamW.
Design notes:
- Inputs are padded to config.max_seq_len inside forward and logits sliced
back, so every training batch presents one static shape (compile/cudagraph
friendly, and uniform kernel shapes even in eager).
- Attention mask gets an iden... | {
"id": "3bfc2977-d62e-44d7-8d16-c42b2bb5504b",
"created_at": "2026-08-06 19:50:08.081438+00:00",
"db_md5": "ef7d67d872a65aa96e81ca53991ab27d",
"submitter": "Jordan Rubin",
"github_login": "jordanrubin",
"run_id": "fe7710a7-4f2a-4a2d-bc32-02f1c0136873",
"tier": "easy",
"dataset_id": "e2",
"status": "s... | {
"score": {
"mean_loss": 3.2833896725763547,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.24374999571591616
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 5.320267186070707,
"example_count": 300,... |
3c1578d8-1796-403f-af92-b6a3c9940d8b | easy | lzy54 | 2026-08-26 07:13:12.033901+00:00 | succeeded | 7bf30e19d90e7f6cadb4ad67d0b2cd51634a4fd2d29a507be6f840a0dd78f8c1 | 3,628 | null | """Basic single-pass Transformer with PyTorch AdamW."""
from __future__ import annotations
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
assert_model_state,
)
D_MODEL = 128
NUM_HEADS = 4
... | {
"id": "3c1578d8-1796-403f-af92-b6a3c9940d8b",
"created_at": "2026-08-26 07:13:12.033901+00:00",
"db_md5": "eaac69fe345f82973fcf392c504a40ff",
"submitter": "lzy54",
"github_login": "lzy54",
"run_id": "6fb8f1e8-4d2f-47d3-abce-d0a97ed8429c",
"tier": "easy",
"dataset_id": "e6",
"status": "succeeded",
... | {
"score": {
"mean_loss": 2.9133031964302063,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.02860360499471426
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 3.8673574924468994,
"example_count": 60,... |
3c15c404-39aa-4bd7-a562-d3c2336b3975 | easy | khushidahi | 2026-08-17 13:29:57.910448+00:00 | succeeded | a962f9c9a43b1dd8b9b9b108d1beb99933b68b76968a4458c51028924b8e77e3 | 29,880 | null | """R191: task-aligned Tiny Recursive Model with answer/workspace states.
This is a learned architecture, not an arithmetic program. Decimal fields are
rearranged into categorical tapes, but no tensor is converted to a native
number and no multiplication, reduction, quotient, carry, or task transition
is specified. A... | {
"id": "3c15c404-39aa-4bd7-a562-d3c2336b3975",
"created_at": "2026-08-17 13:29:57.910448+00:00",
"db_md5": "33596e8ca3a39eb7faabdadd8910d628",
"submitter": "khushidahi",
"github_login": "khushidahi",
"run_id": "3a2810e3-0c81-4234-9a4c-e0d6c7e8d5b8",
"tier": "easy",
"dataset_id": "e5",
"status": "succ... | {
"score": {
"mean_loss": 2.5456453737790876,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.009166666691501936
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.5177576178392487,
"example_count": 60... |
3c268067-7e79-4c97-b902-151a58eb1329 | easy | sapient-sapiens | 2026-08-16 17:09:13.392539+00:00 | succeeded | 1486a457a76b551687bbdcfb359ae9b5b976be18b3d0b7ffe1e700a6b19ffafa | 22,026 | null | """E6 generalization candidate: width-1024, d_h=16, 10-block ACT64.
Architecture changes relative to the hosted ACT max-4 source:
- selectable recurrence depths are 1 through 64;
- canonical cumulative-mass halting genuinely removes halted examples from
later recurrent applications;
- the T-field representation pre... | {
"id": "3c268067-7e79-4c97-b902-151a58eb1329",
"created_at": "2026-08-16 17:09:13.392539+00:00",
"db_md5": "84af624ac7acad7fea2166288ec2123d",
"submitter": "Ertondy",
"github_login": "sapient-sapiens",
"run_id": "dc463a52-2124-44a1-a3f5-f4e313794ea1",
"tier": "easy",
"dataset_id": "e6",
"status": "su... | {
"score": {
"mean_loss": 5.893312699174228,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0770270315980589
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 5.305182456970215,
"example_count": 60,
... |
3c274379-5bcf-43b7-af96-ff5e9bf90fa0 | easy | DDanlov | 2026-08-10 09:33:27.085951+00:00 | failed | 4d8f303e5e4299af6e062fcaa87064fc9152edcaef08a266fe180281b1991de0 | 15,742 | null | from __future__ import annotations
import sys
import os
import math
import time
import torch
import torch.nn as nn
import torch.nn.functional as F
from torch.optim.optimizer import Optimizer
VOCAB_SIZE = 17
DIGIT_OFFSET = 7
class RMSNorm(nn.Module):
def __init__(self, dim: int, eps: float = 1e-6):
super... | {
"id": "3c274379-5bcf-43b7-af96-ff5e9bf90fa0",
"created_at": "2026-08-10 09:33:27.085951+00:00",
"db_md5": "73ffb4bb30b4347fade40531b4b14cd8",
"submitter": "DDanlov",
"github_login": "DDanlov",
"run_id": "542a22d4-b43e-4c59-a0d0-1cb9ae03822c",
"tier": "easy",
"dataset_id": "e1",
"status": "failed",
... | null
|
3c288d78-a02f-43f0-b723-42a4cc2662aa | easy | erdavis0 | 2026-08-29 01:19:13.333611+00:00 | succeeded | c30171170d2368108bbd1f8de018dfbd1d387e0e28dec2d543ae91ea35b7ad59 | 20,136 | null | """Curvature table with a GPU-resident cached modulus-relative residual."""
from __future__ import annotations
import math
import time
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
TokenLos... | {
"id": "3c288d78-a02f-43f0-b723-42a4cc2662aa",
"created_at": "2026-08-29 01:19:13.333611+00:00",
"db_md5": "efb4da441ead29fd5826581673cba669",
"submitter": "Ethan",
"github_login": "erdavis0",
"run_id": "cb259e08-8a03-4968-8ec4-3c13059e6f4c",
"tier": "easy",
"dataset_id": "e5",
"status": "succeeded",... | {
"score": {
"mean_loss": 2.9922684936817356,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.04125000001241763
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 3.023970143142837,
"example_count": 600,... |
3c2c13fe-8c4b-4e8f-a9f6-c260f06d56d7 | easy | DDanlov | 2026-08-29 14:45:51.382882+00:00 | succeeded | b90ab8391b38b2c8b512b141c24b25a4d48fa6a79fa9ef5ae869630daf9403ba | 26,468 | null | """
trial_f05_l5_d768_warmup65.py
F05: L=5, D=768, K_max=36, progress_thresh=0.65, lr=0.0035, update_freq=5
"""
import math
import time
import contextlib
from dataclasses import dataclass
from typing import Optional, Tuple, Dict, Any, List, Union
import torch
import torch.nn as nn
import torch.nn.functional as F
try:... | {
"id": "3c2c13fe-8c4b-4e8f-a9f6-c260f06d56d7",
"created_at": "2026-08-29 14:45:51.382882+00:00",
"db_md5": "4749234a876366f48976080a1c560d4e",
"submitter": "DDanlov",
"github_login": "DDanlov",
"run_id": "cdd4dc18-2727-4153-a703-1da3e63165ec",
"tier": "easy",
"dataset_id": "e5",
"status": "succeeded"... | {
"score": {
"mean_loss": 2.286589029334813,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0141666666790843
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.288447843137878,
"example_count": 600,
... |
3c2d2246-b111-42a5-931e-987a28290654 | easy | yashkant | 2026-08-31 19:24:49.284739+00:00 | succeeded | adc2326f3709c88141cb539694e54735fac514839d98deb598caddf621c78a99 | 106,003 | null | """One Layer Deeper submission -- a learned register machine.
Architecture (all outputs produced by learned parameters; no task arithmetic on the answering path):
a bank of R=6 digit registers (W=16 positions each) carried as DISCRETE tokens between micro-steps: five
data registers and one COUNTDOWN register holding t... | {
"id": "3c2d2246-b111-42a5-931e-987a28290654",
"created_at": "2026-08-31 19:24:49.284739+00:00",
"db_md5": "17dba5fed34dfa2174d34ffdc9d632ee",
"submitter": "Yash Kant",
"github_login": "yashkant",
"run_id": "85149cc9-ca12-4f76-9a01-2e434efde683",
"tier": "easy",
"dataset_id": "e4",
"status": "succeed... | {
"score": {
"mean_loss": 5.337064150776815,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 5.318246069510422,
"example_count": 1200,
"exac... |
3c2fc019-beb3-44aa-840f-114272e0ffc5 | easy | c-nnr | 2026-08-04 11:49:59.673670+00:00 | succeeded | c70e3e81a76b03938003166e77fbbd3f63cd3d12fb3090a443c2a91458cebb27 | 4,294 | null | """One Layer Deeper experiment: Baseline with equal per-sequence cross-entropy weighting."""
from __future__ import annotations
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
TokenLossBatch,
... | {
"id": "3c2fc019-beb3-44aa-840f-114272e0ffc5",
"created_at": "2026-08-04 11:49:59.673670+00:00",
"db_md5": "d9e5be2fd8542fbe03f2a4b94d732db6",
"submitter": "c-nnr",
"github_login": "c-nnr",
"run_id": "c1b47bda-01c7-4926-a913-5a59d61719f4",
"tier": "easy",
"dataset_id": "e1",
"status": "succeeded",
... | {
"score": {
"mean_loss": 3.2277305126190186,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.014999999664723873
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 4.296428680419922,
"example_count": 100... |
3c37fd64-11de-41c4-88aa-b06b029a2e5b | easy | viridale | 2026-08-13 10:47:32.099730+00:00 | succeeded | 53a4adbe389695841b9e315552e6267df423844c6f839d348f40508ccdf671c9 | 26,663 | null | """Shared latent-action Cayley atlas with completed-path supervision.
hypothesis: deep terminal rows should supervise shallow learned power maps directly,
while a shared depth-conditioned latent action and adjacent/balanced Cayley paths
remove independent-head root freedom without a decimal-distribution bottle... | {
"id": "3c37fd64-11de-41c4-88aa-b06b029a2e5b",
"created_at": "2026-08-13 10:47:32.099730+00:00",
"db_md5": "b6f1f30fd5fc4ff1fc009c38e8b5df12",
"submitter": "priormancer",
"github_login": "viridale",
"run_id": "f4e525a3-d051-4aa1-a303-d4e57139dfd1",
"tier": "easy",
"dataset_id": "e3",
"status": "succe... | {
"score": {
"mean_loss": 2.132656216621399,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.004999999888241291
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.133573055267334,
"example_count": 800,... |
3c388cb5-92a1-4e30-b9b1-99461c94280c | easy | jordanrubin | 2026-08-02 18:06:55.346034+00:00 | succeeded | 5673d4c12fa46edd2a2b84ca528c969ec34d5e313e6f1f7e3784c935f5e1e388 | 13,992 | null | """Bidirectional transformer, D=256 x 3 blocks, grokking-tuned AdamW.
Design notes:
- Inputs are padded to config.max_seq_len inside forward and logits sliced
back, so every training batch presents one static shape (compile/cudagraph
friendly, and uniform kernel shapes even in eager).
- Attention mask gets an iden... | {
"id": "3c388cb5-92a1-4e30-b9b1-99461c94280c",
"created_at": "2026-08-02 18:06:55.346034+00:00",
"db_md5": "3ed90663304f1de06b838c5aeee9e9b0",
"submitter": "Jordan Rubin",
"github_login": "jordanrubin",
"run_id": "7beafbbd-c316-479f-8b4a-dde5d4815b28",
"tier": "easy",
"dataset_id": "e1",
"status": "s... | {
"score": {
"mean_loss": 2.3549550669539885,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0516666645805041
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.7379718707568608,
"example_count": 100,... |
3c41018e-0c61-428a-a966-f9d9b3a70c8f | easy | viridale | 2026-08-08 20:16:25.122722+00:00 | succeeded | 51c99b81541af1aecd5733b0b5a9cc83215f35f8f462c0782bf0794d91747ddf | 12,243 | null | """Shared learned bilinear algebra over canonical integer coordinates.
hypothesis: v372 distilled 250 exact table rows yet left reserved T1 unchanged at 7/38,
proving a generic relation MLP does not discover the operation. Make self-product
native: shared numeric coordinates enter two independently learned fa... | {
"id": "3c41018e-0c61-428a-a966-f9d9b3a70c8f",
"created_at": "2026-08-08 20:16:25.122722+00:00",
"db_md5": "0c0189daf5e4e3018ed04e1b06083c75",
"submitter": "priormancer",
"github_login": "viridale",
"run_id": "c1fff143-a8f0-4d38-af7e-9e90915d1feb",
"tier": "easy",
"dataset_id": "e1",
"status": "succe... | {
"score": {
"mean_loss": 1.9622805714607239,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.37999999336898327
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.9716901779174805,
"example_count": 100... |
3c432055-4ac1-446a-8c12-24f23be48395 | easy | Dandandan | 2026-08-07 08:59:16.134409+00:00 | succeeded | f86e91b6e635dfbf4e48787d264ced72273deb8ee17f473efd4b8d61781bcda7 | 216,832 | null | """Standalone generic neural-ALU router; mechanically bundled."""
from __future__ import annotations
import types as _types
def _make_shell():
import math
import time
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
ModelSpec,
O... | {
"id": "3c432055-4ac1-446a-8c12-24f23be48395",
"created_at": "2026-08-07 08:59:16.134409+00:00",
"db_md5": "9dc626d293f39aa0d78deba4f15b5d5f",
"submitter": "Daniël Heres",
"github_login": "Dandandan",
"run_id": "85f3db0b-633f-430e-a82c-ebcd19ed9ca5",
"tier": "easy",
"dataset_id": "e1",
"status": "suc... | {
"score": {
"mean_loss": 1.1271511084487429e-05,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 1.0
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 1.1280336366326082e-05,
"example_count": 100,
... |
3c437cf4-cee9-494f-b6a0-b731aaa6ab38 | easy | yashkant | 2026-08-23 20:38:34.243062+00:00 | failed | 35db76f466990cd8e74833393b591825bd4ccfdd7634c8698c959b560bb138b1 | 11,550 | null | """One Layer Deeper submission: CRT residue-register machine with a learned squaring
transition and a Fourier "clock" bilinear reconstruction readout.
Idea (end-to-end learned; nothing about the task is hard-coded):
* State = per-modulus residue registers over a fixed bank of small primes. The input
integer x is... | {
"id": "3c437cf4-cee9-494f-b6a0-b731aaa6ab38",
"created_at": "2026-08-23 20:38:34.243062+00:00",
"db_md5": "54b80a60a8aca82369ec363db1993af0",
"submitter": "Yash Kant",
"github_login": "yashkant",
"run_id": "83a1f7bf-1d92-45f5-90b3-91b1bfd7b0df",
"tier": "easy",
"dataset_id": "e1",
"status": "failed"... | null
|
3c44d344-bea8-4f78-b7d3-fdbffe84be9e | easy | yashkant | 2026-08-16 00:54:09.201972+00:00 | succeeded | f8d6a10f5d464a51f9925de8997cd6639f3af95fbfc41d2c680460510463778f | 36,180 | null | """Round1763 residual-conditioned shallow/deep Dykstra mixer.
Every parsed public-T cycle exposes the state after its first shared A
projection (the 1*T shallow readout) and after the full shared A-B-A cycle (the
3*T deep readout). Differentiable state-update energies plus a terminal/mean
convergence ratio drive a tin... | {
"id": "3c44d344-bea8-4f78-b7d3-fdbffe84be9e",
"created_at": "2026-08-16 00:54:09.201972+00:00",
"db_md5": "076757e2ade7d2d10a5818a56118d4c5",
"submitter": "Yash Kant",
"github_login": "yashkant",
"run_id": "1bf3eebd-2d09-4554-8e78-e876507405c4",
"tier": "easy",
"dataset_id": "e5",
"status": "succeed... | {
"score": {
"mean_loss": 2.1525719499284186,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.010833333345750967
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.1578440270616333,
"example_count": 60... |
3c4556e2-3a88-4828-a351-74a2f9389a1a | easy | oupadhyay | 2026-08-08 10:57:19.381047+00:00 | succeeded | e8fe9f06f85d2c0ca036b6899185a7bf40819c239adeb45a5dcd38ac54f16357 | 11,453 | null | """Shared categorical digit transition with a generic DNC working memory."""
from __future__ import annotations
import math
import torch
from torch import Tensor, nn
import torch.nn.functional as F
from benchmark import ModelSpec, OptimizerBundle, OptimizerSpec, Submission, assert_model_state
PAD, N_MARK, X_MARK, T_... | {
"id": "3c4556e2-3a88-4828-a351-74a2f9389a1a",
"created_at": "2026-08-08 10:57:19.381047+00:00",
"db_md5": "2f4ed946c4de590fd1a3fe1e14b58254",
"submitter": "Ojasw Upadhyay",
"github_login": "oupadhyay",
"run_id": "a98cfba9-566b-435c-953e-be7f4270731d",
"tier": "easy",
"dataset_id": "e5",
"status": "s... | {
"score": {
"mean_loss": 2.1506236792229267,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.00375
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.1510134213708443,
"example_count": 600,
... |
3c4696c8-a214-4c27-9c87-4107462465b5 | easy | DDanlov | 2026-08-18 10:53:35.605491+00:00 | succeeded | 9e2e886a7f72218c6f6788250c79c4edcba3908cfb09a02021dbdad2941008ba | 23,584 | null | """
Official Submission for One Layer Deeper Challenge
Architecture: 4-Layer Recurrent Core Pure No-Norm No-Skip DKS Reasoning Model:
- 4 Distinct Pure No-Norm No-Skip Transformer Layers in the Recurrent Core: L1 -> L2 -> L3 -> L4
- Unrolled T times for serial modular squaring
- dim=512, 16 heads, d_ff=1024, RohanShamp... | {
"id": "3c4696c8-a214-4c27-9c87-4107462465b5",
"created_at": "2026-08-18 10:53:35.605491+00:00",
"db_md5": "d6174d2a9211082098f18de894eca987",
"submitter": "DDanlov",
"github_login": "DDanlov",
"run_id": "97d04678-d20f-43fd-aa25-d3d0bb522d17",
"tier": "easy",
"dataset_id": "e3",
"status": "succeeded"... | {
"score": {
"mean_loss": 2.3104383395965256,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.006875
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.310266181205971,
"example_count": 800,
... |
3c4986c0-453a-489c-803e-0dbd9303671f | easy | 0Chris5R | 2026-08-07 12:41:46.381940+00:00 | succeeded | 16a4a885220ed5c79b80bc977dfa1218767441f8b8004e369991221a12acb78d | 20,838 | null | """Shared-digit Neural GPU with a compute-matched operator curriculum.
The model aligns decimal digits by significance, forms a learned low-rank
second-order feature map of the current state, and evolves a two-dimensional
workspace with the original convolutional-GRU equations. Each arithmetic
transition applies the ... | {
"id": "3c4986c0-453a-489c-803e-0dbd9303671f",
"created_at": "2026-08-07 12:41:46.381940+00:00",
"db_md5": "a135532c71a0604cf839d4cc3e79cbc9",
"submitter": "Chris ",
"github_login": "0Chris5R",
"run_id": "2a250fd2-e179-444b-8bb1-f9867620198c",
"tier": "easy",
"dataset_id": "e5",
"status": "succeeded"... | {
"score": {
"mean_loss": 2.987190975406265,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.032083333370586234
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.964620570965412,
"example_count": 600,... |
3c4a9998-ab95-4de8-97e4-4b9e671c0584 | easy | DDanlov | 2026-08-13 22:38:42.985778+00:00 | failed | 654bcba4cea296daf285211d1b0d76aeccbd2a9385f9b8f0239cc1c4836874b8 | 22,145 | null | import torch
import torch.nn as nn
import torch.nn.functional as F
import math
from dataclasses import dataclass
from typing import Optional
from torch.optim import Optimizer
try:
from client import Submission, assert_model_state, OptimizerBundle
except ImportError:
Submission = None
assert_model_state = N... | {
"id": "3c4a9998-ab95-4de8-97e4-4b9e671c0584",
"created_at": "2026-08-13 22:38:42.985778+00:00",
"db_md5": "fec3750d33d011c885790a1ecc498900",
"submitter": "DDanlov",
"github_login": "DDanlov",
"run_id": "d58df4e9-89cd-4257-b4d9-4dec3360f945",
"tier": "easy",
"dataset_id": "e3",
"status": "failed",
... | null
|
3c50339e-b7e0-482e-8453-5a57fa30b086 | easy | 0Chris5R | 2026-08-02 15:28:35.233311+00:00 | failed | 2dcbbc111f3584bf22fdf3b169b32ce38e4cf22651039ebaa61eda8750c253e4 | 44,725 | null | from __future__ import annotations
from dataclasses import asdict, dataclass
import math
import time
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
assert_model_state,
)
SUBMISSION_STARTED_... | {
"id": "3c50339e-b7e0-482e-8453-5a57fa30b086",
"created_at": "2026-08-02 15:28:35.233311+00:00",
"db_md5": "887da97efd1780b4b131c8ebba1c0996",
"submitter": "Chris ",
"github_login": "0Chris5R",
"run_id": "beb531d9-8cf2-4fc4-b740-c0dc3c1ab27f",
"tier": "easy",
"dataset_id": "e5",
"status": "failed",
... | null
|
3c51d37b-053b-4c98-a487-0aef46e4fc75 | easy | DDanlov | 2026-08-14 07:46:44.299816+00:00 | succeeded | d99c6ad82a361cc7d0d597beb98d01ba5af5c9a0a34928a39c7768e5a46bcc9a | 21,975 | null | import math
from typing import Optional
import torch
import torch.nn as nn
import torch.nn.functional as F
from torch.optim import Optimizer
try:
from benchmark import Submission, assert_model_state, OptimizerBundle
except ImportError:
try:
from client import Submission, assert_model_state, OptimizerBu... | {
"id": "3c51d37b-053b-4c98-a487-0aef46e4fc75",
"created_at": "2026-08-14 07:46:44.299816+00:00",
"db_md5": "6b69f6e680bce977975601e0f42f358d",
"submitter": "DDanlov",
"github_login": "DDanlov",
"run_id": "404efb1d-708d-4dff-9e3b-60ed1209a837",
"tier": "easy",
"dataset_id": "e3",
"status": "succeeded"... | {
"score": {
"mean_loss": 2.7911358489094145,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.018125000000000002
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.7913867589752366,
"example_count": 80... |
3c534669-93ef-4123-b612-ba62cd415963 | easy | DDanlov | 2026-08-11 20:59:05.234024+00:00 | succeeded | 9d850cd083e0ec0cd84d0d6c9ce21a4ecfc4a056dedc41de01945e1761548eb8 | 19,621 | null | from __future__ import annotations
import sys
import os
import math
import time
import torch
import torch.nn as nn
import torch.nn.functional as F
from torch.optim.optimizer import Optimizer
VOCAB_SIZE = 17
DIGIT_OFFSET = 7
def init_cutting_the_skip(module):
"""SVD-based orthogonal initialization to ensure initia... | {
"id": "3c534669-93ef-4123-b612-ba62cd415963",
"created_at": "2026-08-11 20:59:05.234024+00:00",
"db_md5": "46d46ecd1a6bb6857435da654b52c662",
"submitter": "DDanlov",
"github_login": "DDanlov",
"run_id": "9da445a5-c726-4768-b4e5-ebd0d618f7be",
"tier": "easy",
"dataset_id": "e1",
"status": "succeeded"... | {
"score": {
"mean_loss": 419.5556786643268,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 329.7703480507011,
"example_count": 100,
"exact... |
3c569a26-df7f-48c5-bf90-5b07b6f9c201 | easy | Cree0618 | 2026-08-29 10:58:00.488632+00:00 | succeeded | 0f246b0c878ef87b0a55b8915d4e122f823bdebcd48bd5d6330e0de3fa6851a2 | 4,345 | null | """E019: true single-pass Transformer with an untied output head.
Hypothesis: E007's four repeated uses of one shared block spend most of the
fixed training budget on a recurrence that aids memorization but has shown no
transfer. A single block should allow more optimizer updates, while the untied
head keeps E007's re... | {
"id": "3c569a26-df7f-48c5-bf90-5b07b6f9c201",
"created_at": "2026-08-29 10:58:00.488632+00:00",
"db_md5": "c8659299b474b2f8361cbf14566e5092",
"submitter": "Cree0618",
"github_login": "Cree0618",
"run_id": "3e17546c-7df1-464c-bd9e-4d2d75187807",
"tier": "easy",
"dataset_id": "e3",
"status": "succeede... | {
"score": {
"mean_loss": 3.0914206379546467,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.005000000009313226
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 3.422409938672635,
"example_count": 800... |
3c59ac0b-459e-47b7-a936-9839905ee435 | easy | shirvani-jr | 2026-08-29 00:35:47.990698+00:00 | succeeded | 042eb09079dba2b92f40d9887b2e888f0aff9068d42e7814be28260ea9128814 | 49,054 | null | """Adaptive learned discrete recurrent fabric for One Layer Deeper.
A gradient-trained, operation-free recurrent computational machine. The
model learns, from the competition's endpoint supervision: how to parse the
prompt into a bank of categorical digit registers, a reusable transition
circuit applied serially, a l... | {
"id": "3c59ac0b-459e-47b7-a936-9839905ee435",
"created_at": "2026-08-29 00:35:47.990698+00:00",
"db_md5": "d86f78b0f3ee82b6472811f319120edb",
"submitter": "Ali",
"github_login": "shirvani-jr",
"run_id": "241bbba2-0662-48ed-90bc-b991985a3deb",
"tier": "easy",
"dataset_id": "e1",
"status": "succeeded"... | {
"score": {
"mean_loss": 2.1587535738945007,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.19666667096316814
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.473297119140625,
"example_count": 100,... |
3c5b2939-040d-4a00-8e91-d6f05c09604a | easy | yunjiangster | 2026-08-27 21:49:37.200378+00:00 | succeeded | 1cd38f959f0dc51413f544c0bdb921cb880626e9cad3200136a56494d5c319a8 | 14,801 | null | """Generic learned-program probe for One Layer Deeper.
A randomly initialized categorical parameter selects one program from a generic
accumulator-machine DSL. Programs are generated mechanically from four generic
instructions over the public scalar input fields; no modular-squaring expression
or arithmetic trace is ... | {
"id": "3c5b2939-040d-4a00-8e91-d6f05c09604a",
"created_at": "2026-08-27 21:49:37.200378+00:00",
"db_md5": "b2a31ee68cfe204246c2a2e8d3b276ba",
"submitter": "Yunjiang Jiang",
"github_login": "yunjiangster",
"run_id": "55312065-0bec-4a86-a341-86dffd7b5cfe",
"tier": "easy",
"dataset_id": "e2",
"status":... | {
"score": {
"mean_loss": 0.004424781072884798,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 1.0
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 0.004424781072884798,
"example_count": 300,
... |
3c5b567a-a485-4d07-93fa-66c0ea612020 | easy | erdavis0 | 2026-08-25 02:24:50.999967+00:00 | succeeded | eb53af5bf05220180cc36fcae77167a9bf953b2d302dcb51899a4361ba4b15f9 | 27,217 | null | """Reversible augmented-state transducer over aligned learned digit tapes.
Public delimiter syntax only routes opaque token categories into distinct,
right-aligned N, X, and T tapes. Learned local cells compile separate
interaction, carry, and modulus-conditioned workspaces. N is isolated from
interaction and carry ... | {
"id": "3c5b567a-a485-4d07-93fa-66c0ea612020",
"created_at": "2026-08-25 02:24:50.999967+00:00",
"db_md5": "ec284dc76774dfac34157df0311358c6",
"submitter": "Ethan",
"github_login": "erdavis0",
"run_id": "79f1c31d-19eb-411a-af55-978d820a0522",
"tier": "easy",
"dataset_id": "e5",
"status": "succeeded",... | {
"score": {
"mean_loss": 4.77380286944991,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.015000000037252903
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 4.70810400317068,
"example_count": 600,
... |
3c603fd6-70ec-492b-a09a-174712bb8596 | easy | DDanlov | 2026-08-28 22:14:18.930006+00:00 | succeeded | e08f2fed5e545fc90c7866705897de407dcbb9bfb892bb99f747bf41a86f643e | 15,428 | null | """
Trial 38: Model D: D=384, H=6, M=4, u=2 (T=8)
"""
import math
import torch
import torch.nn as nn
import torch.nn.functional as F
from dataclasses import dataclass
from typing import Optional, Tuple, Dict, Any, List, Union
try:
from benchmark.api import ModelSpec, OptimizerBundle, OptimizerSpec, Submission, To... | {
"id": "3c603fd6-70ec-492b-a09a-174712bb8596",
"created_at": "2026-08-28 22:14:18.930006+00:00",
"db_md5": "27dfc55fe7b3319741199174b6bba5d8",
"submitter": "DDanlov",
"github_login": "DDanlov",
"run_id": "57f4d1aa-9574-48d6-8901-f2f67adbab3c",
"tier": "easy",
"dataset_id": "e5",
"status": "succeeded"... | {
"score": {
"mean_loss": 2.3392674891520526,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.002916666666666667
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.3396404479116604,
"example_count": 60... |
3c6e10b2-e099-40d0-9efa-db2380e4752e | easy | viridale | 2026-08-25 10:40:22.988758+00:00 | succeeded | 09e3de4c5c921bcbba881c44d1a24910d6a90bcbba8aba2eee387035bc960bfa | 23,088 | null | """Constraint-posterior bilinear Transformer program.
hypothesis: v890 proved even a tiny explicit prior biases the T64 categorical
boundary; retaining the lower-dimensional learned global-frame path but removing
both penalties should preserve exact contextual coefficients while optimizer geometry
still fa... | {
"id": "3c6e10b2-e099-40d0-9efa-db2380e4752e",
"created_at": "2026-08-25 10:40:22.988758+00:00",
"db_md5": "de994e5e701f5aa98d961b926e293255",
"submitter": "priormancer",
"github_login": "viridale",
"run_id": "9ba0359c-6b02-494b-805e-8ef676f61e50",
"tier": "easy",
"dataset_id": "e6",
"status": "succe... | {
"score": {
"mean_loss": 0.0,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 1.0
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 0.0,
"example_count": 60,
"exact_accuracy": 1.0,
"c... |
3c70ae74-422a-4cd5-8b1a-d41d35811f0a | easy | erdavis0 | 2026-08-21 00:49:25.397511+00:00 | succeeded | 336483c3d53caddd4d0b9404b43729690356ba8c26360b530b11e1dd341476c7 | 32,848 | null | """Learned global microprogram over a sealed categorical tape."""
from __future__ import annotations
from dataclasses import dataclass
from types import SimpleNamespace
import torch
import torch.nn.functional as F
from torch import nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
... | {
"id": "3c70ae74-422a-4cd5-8b1a-d41d35811f0a",
"created_at": "2026-08-21 00:49:25.397511+00:00",
"db_md5": "9029bd2bb72ac7670a520c5b4227acc0",
"submitter": "Ethan",
"github_login": "erdavis0",
"run_id": "e0615abf-df29-424a-a55e-a138b4030e0a",
"tier": "easy",
"dataset_id": "e3",
"status": "succeeded",... | {
"score": {
"mean_loss": 2.1837577266819066,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.010624999999999999
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.182765938125341,
"example_count": 800... |
3c77706f-cdde-4b71-b249-fe75a37429cc | easy | lpbb | 2026-08-07 12:00:01.506352+00:00 | succeeded | 8034c2375475d0fa182a7c2c099d37232193777b5dbd9ee8069b6ed805aa8808 | 4,326 | null | """Basic single-pass Transformer with PyTorch AdamW."""
from __future__ import annotations
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
assert_model_state,
)
# Literal hyperparameter dict... | {
"id": "3c77706f-cdde-4b71-b249-fe75a37429cc",
"created_at": "2026-08-07 12:00:01.506352+00:00",
"db_md5": "752b8c4b53b37c5fa3e276f4404e3d02",
"submitter": "Lpbb",
"github_login": "lpbb",
"run_id": "965a40d6-1a32-45a7-addc-47ddbde7ebc5",
"tier": "easy",
"dataset_id": "e3",
"status": "succeeded",
"s... | {
"score": {
"mean_loss": 3.794157300589162,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.007500000009313226
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 4.907275569805038,
"example_count": 800,... |
3c7860a7-4715-471e-a13a-c463090ceeaa | easy | erdavis0 | 2026-08-24 16:20:44.676582+00:00 | succeeded | 1d143e0993d68891261502d435cd3b70f96f2c4ceda9481c9fce0012b270a3e9 | 12,744 | null | """Generic recurrent mixer in a learned tensor-product coordinate system.
Raw token and position features are contextualized once. A learned analysis
frame maps latent state and prompt condition into independent coordinates,
their complete outer product is contracted through a learned three-way core,
and a learned sy... | {
"id": "3c7860a7-4715-471e-a13a-c463090ceeaa",
"created_at": "2026-08-24 16:20:44.676582+00:00",
"db_md5": "16dfe6c08c51bc02d57459e3b0a0fc14",
"submitter": "Ethan",
"github_login": "erdavis0",
"run_id": "5eaa233d-a9fd-49ce-a83e-620c4914854d",
"tier": "easy",
"dataset_id": "e6",
"status": "succeeded",... | {
"score": {
"mean_loss": 6.5339195728302,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.053603606298565865
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 5.760258674621582,
"example_count": 60,
... |
3c7e7557-dc9d-4042-883d-0e15e5682b0c | easy | shreyash-chonkie | 2026-09-01 03:21:07.677327+00:00 | failed | bd16c29015b34ae68cf817fec2b39e756da156ed097ff67633592190eb5f2a78 | 20,008 | null | """Learned token scanner feeding a random recurrent path reservoir."""
from __future__ import annotations
from itertools import permutations
import torch
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
TokenLossBatch,
assert_model_s... | {
"id": "3c7e7557-dc9d-4042-883d-0e15e5682b0c",
"created_at": "2026-09-01 03:21:07.677327+00:00",
"db_md5": "c52a7411490c902802878dfb66f3fe6b",
"submitter": "Shreyash",
"github_login": "shreyash-chonkie",
"run_id": "a2fb9d66-e4bc-4917-80a9-31f0700f3ac9",
"tier": "easy",
"dataset_id": "e3",
"status": "... | null
|
3c86b243-9788-4af7-9f53-7879a8ca2576 | easy | yashkant | 2026-08-15 14:02:03.016526+00:00 | succeeded | 5835c5090906a31e0f11494768841784c59e00fe4325e2d747ae6693c783374e | 32,226 | null | """Round1232 parsed-T allocation crossover on the y/z MinMax transition.
The complete Round1220 answer/latent MinMax candidate remains unchanged except
for the public-T-to-macro allocation. Separate answer y and latent z states,
immutable x reinjection, both MinMax updates per macro, and full BPTT remain
intact. MINMA... | {
"id": "3c86b243-9788-4af7-9f53-7879a8ca2576",
"created_at": "2026-08-15 14:02:03.016526+00:00",
"db_md5": "25e502d53e86da62dc76f02456a9b310",
"submitter": "Yash Kant",
"github_login": "yashkant",
"run_id": "a3b11c61-7a49-4d3c-ac3a-ae84dda63e08",
"tier": "easy",
"dataset_id": "e2",
"status": "succeed... | {
"score": {
"mean_loss": 3.069519579410553,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.2952083572745323
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 4.574514389038086,
"example_count": 300,
... |
3c92c78f-ff96-4eec-9bb8-5561828805c5 | easy | DDanlov | 2026-08-09 00:05:15.380824+00:00 | failed | cb443ebefc074d283d851d1a9ced0f7e93140fd26168eaeaaeb284e271522842 | 15,981 | null | from __future__ import annotations
import sys
import os
import math
import time
import torch
import torch.nn as nn
import torch.nn.functional as F
from torch.optim.optimizer import Optimizer
VOCAB_SIZE = 17
DIGIT_OFFSET = 7
class RohanShampoo(Optimizer):
"""Self-contained RohanShampoo optimizer with eigenvalue ma... | {
"id": "3c92c78f-ff96-4eec-9bb8-5561828805c5",
"created_at": "2026-08-09 00:05:15.380824+00:00",
"db_md5": "ce45fa7ef2e01ebc15bf99ed070c5256",
"submitter": "DDanlov",
"github_login": "DDanlov",
"run_id": "603d50b8-746e-43ac-b3ec-363d924ec3c6",
"tier": "easy",
"dataset_id": "e1",
"status": "failed",
... | null
|
3c93a212-9b45-4c63-8729-fe951cd224dc | easy | falloficaruss | 2026-08-07 09:21:03.687704+00:00 | succeeded | ce74b1ceae5e1bd38ab042bde46a8eb4f3e40f5090b9bb06dffbedc8daa47657 | 23,322 | null | """T-gated latent digit square-iterator v2 for Medium m1.
Abacus PE, LSD digit registers, hard step<T gate, x0 inject + FiLM(step,T,rem),
TRM refine aux on final labels, AdamW. No adaptive-depth mechanism and no
intermediate-residue targets.
"""
from __future__ import annotations
import math
import torch
import tor... | {
"id": "3c93a212-9b45-4c63-8729-fe951cd224dc",
"created_at": "2026-08-07 09:21:03.687704+00:00",
"db_md5": "cdd610c698a697b078ff6a9d8ee5e59c",
"submitter": "Abhishek Shinde",
"github_login": "falloficaruss",
"run_id": "144e117b-e20c-48ea-83b1-b9a6d1dc204b",
"tier": "easy",
"dataset_id": "e1",
"status... | {
"score": {
"mean_loss": 1.846206247806549,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.044999999925494194
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 1.7985535860061646,
"example_count": 100... |
3c9402ad-8387-4525-b230-65014b586019 | easy | DESU-CLUB | 2026-08-24 10:05:37.414555+00:00 | succeeded | ce76301768085d521fd3d1e0236f790a91f21d61abe0cbe58443b6724f717679 | 7,898 | null | """Huginn-style looped Transformer whose iteration count is read from the T token.
prelude -> [core applied T times, input re-injected each step] -> coda
Forked from experiments/_baselines/adamw_transformer.py; the Block, RMSNorm and
optimizer are unchanged so that the only varied axis is the depth mechanism.
"""
fr... | {
"id": "3c9402ad-8387-4525-b230-65014b586019",
"created_at": "2026-08-24 10:05:37.414555+00:00",
"db_md5": "68bdc8a08618a83d90c208df60fc77f4",
"submitter": "Low Keng Hoong, Warren",
"github_login": "DESU-CLUB",
"run_id": "e93ce0ae-2dff-4606-9d2e-f7f32eb59a20",
"tier": "easy",
"dataset_id": "e6",
"sta... | {
"score": {
"mean_loss": 6.98771243552639,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.117342346020647
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 5.306178092956543,
"example_count": 60,
... |
3c9ae0ee-2119-4ea6-a29d-1d81ee2fdff4 | easy | shan23chen | 2026-08-08 01:49:58.587351+00:00 | succeeded | 7fba100962e5efa187bd646f9f38531941afe7d3208bde6068f04a222e5f9cb9 | 20,440 | null | """Generic recurrent Transformer with normalized numeric and CRF objectives.
The model is deliberately task-generic: it learns a transition over prompt and
workspace tokens instead of embedding the public modular-squaring algorithm.
A learned linear-chain CRF models dependencies among adjacent output digits;
its endpo... | {
"id": "3c9ae0ee-2119-4ea6-a29d-1d81ee2fdff4",
"created_at": "2026-08-08 01:49:58.587351+00:00",
"db_md5": "cc1680bfcc0e91543a10db4ab76946dc",
"submitter": "Shan Chen",
"github_login": "shan23chen",
"run_id": "43e78136-c69d-462e-8d91-f4cc589f992f",
"tier": "easy",
"dataset_id": "e5",
"status": "succe... | {
"score": {
"mean_loss": 2.844704508781433,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.01000000024214387
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.8294308185577393,
"example_count": 600,... |
3c9ce519-5989-4555-b149-f3a7ab9d2919 | easy | jordanrubin | 2026-08-01 20:15:39.316350+00:00 | succeeded | 2c6adbc91db77091eaa1fdb53b2b15fdc6e05306608a2440be5e6a2aa2591b8a | 18,183 | null | """Looped bidirectional selective-SSM (Mamba-style), pure PyTorch.
The statefulness bet: input-dependent SSM dynamics along the token axis,
with the block weight-tied and applied NUM_LOOPS times in depth with input
injection — recurrence in depth is where the task's T-iteration lives.
The selective scan is computed i... | {
"id": "3c9ce519-5989-4555-b149-f3a7ab9d2919",
"created_at": "2026-08-01 20:15:39.316350+00:00",
"db_md5": "35fec302dbad6ab1de27b5092ede0b6a",
"submitter": "Jordan Rubin",
"github_login": "jordanrubin",
"run_id": "c27e6ad2-5d69-4e2a-8224-51e83e46e56f",
"tier": "easy",
"dataset_id": "e1",
"status": "s... | {
"score": {
"mean_loss": 5.256585295360879,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.036666665847102806
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 5.517359256744385,
"example_count": 100,... |
3ca10c70-fdfd-4613-82c0-b7db0a49be7d | easy | liam-gb | 2026-08-17 02:33:26.222680+00:00 | succeeded | 82632945ab7e0c87d203dfc4e2f107a9e7f49bf180d7e486a194d56b896646c1 | 14,535 | null | """charblock: a from-scratch body whose per-prime step has a known closed-form
solution, trained in the regime where that solution is documented to be found
from random initialisation.
Five ingredients, declared as one ARCHITECTURE+REGIME bundle:
1. Per-prime two-layer blocks with a QUADRATIC activation. For each ban... | {
"id": "3ca10c70-fdfd-4613-82c0-b7db0a49be7d",
"created_at": "2026-08-17 02:33:26.222680+00:00",
"db_md5": "a79f5e7cd40035de9d3d3adb1e10b7b5",
"submitter": "liam-gb",
"github_login": "liam-gb",
"run_id": "63792e9d-c431-49e2-af69-a6c62ffd2e25",
"tier": "easy",
"dataset_id": "e8",
"status": "succeeded"... | {
"score": {
"mean_loss": 2.3792284727096558,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.4233511704951525
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.640991687774658,
"example_count": 85,
... |
3ca37b39-dcda-43b0-bf89-62a9cc2d7f2e | easy | repst | 2026-08-05 21:19:25.589097+00:00 | failed | b4bff7a3f2d6d3483299d49b9027154101979e7fd4b79adde84ee3cee894aaf2 | 13,370 | null | """Soft-routed recurrent neural machine with two evaluation steps."""
from __future__ import annotations
from typing import NamedTuple
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
TokenLos... | {
"id": "3ca37b39-dcda-43b0-bf89-62a9cc2d7f2e",
"created_at": "2026-08-05 21:19:25.589097+00:00",
"db_md5": "cb5437de1c4a42a29396d1a4fc2c1718",
"submitter": "Ryan Epstein",
"github_login": "repst",
"run_id": "a8768253-c8f3-46cc-8a7c-10d0270ddcf5",
"tier": "easy",
"dataset_id": "e5",
"status": "failed"... | null
|
3cab65fc-1406-490f-bf7b-4021232ed58a | easy | KaustubhKumar05 | 2026-08-28 04:57:29.826198+00:00 | succeeded | 721598873c1b37e17fa8987f12fa81202c87610723ab329b1ebc1364f2ea7f1f | 11,140 | null | """op2_abacus -- cell_op2 + ONLY the Abacus random significance offset.
GPT's 2x2 to separate two readings of the S89 ablation. Removing the Fourier digit
features costs -3.69pp, but cell_fix also carries an Abacus random-significance offset,
and BOTH could be doing the same job: not "supplying an arithmetic represent... | {
"id": "3cab65fc-1406-490f-bf7b-4021232ed58a",
"created_at": "2026-08-28 04:57:29.826198+00:00",
"db_md5": "6f5954be43dcc33c08fc61d2a14ae7ab",
"submitter": "koz",
"github_login": "KaustubhKumar05",
"run_id": "a9d65c65-7614-4e25-957e-33243d8dee64",
"tier": "easy",
"dataset_id": "e6",
"status": "succee... | {
"score": {
"mean_loss": 4.313904047012329,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.19954956322908401
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 4.023299694061279,
"example_count": 60,
... |
3caebcf0-1b75-48eb-936d-8b229ebb478f | easy | khushidahi | 2026-08-12 22:01:30.125418+00:00 | failed | ff95a4001f484a47015100a5e5c515c1da0557ddfd561cd54b8bd1f03e16cd9e | 16,711 | null | """R6 structured decimal-workspace model for One Layer Deeper.
The model does not implement multiplication, modular reduction, or the public
recurrence. It uses the public tokenizer structure to build learned, right-
aligned decimal tapes for N, X, and T, then applies a shared neural transition
to a mutable work tape.... | {
"id": "3caebcf0-1b75-48eb-936d-8b229ebb478f",
"created_at": "2026-08-12 22:01:30.125418+00:00",
"db_md5": "1133b04d215f3377979dd431e950e7cb",
"submitter": "khushidahi",
"github_login": "khushidahi",
"run_id": "ff8b1772-0877-408c-ac1a-b4d88d825309",
"tier": "easy",
"dataset_id": "e5",
"status": "fail... | null
|
3cb8da8b-e383-43fe-8c8f-a96f1dc04ce0 | easy | khushidahi | 2026-08-01 21:33:06.523288+00:00 | succeeded | 7a04544ffd807d83bc695b56d573dd9b4419cff8a258bec2c4d609d13bcf4f52 | 4,045 | null | """Basic single-pass Transformer with PyTorch AdamW."""
from __future__ import annotations
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
assert_model_state,
)
D_MODEL = 128
NUM_HEADS = 4
... | {
"id": "3cb8da8b-e383-43fe-8c8f-a96f1dc04ce0",
"created_at": "2026-08-01 21:33:06.523288+00:00",
"db_md5": "1670910d3333b5246604492f0ed98978",
"submitter": "khushidahi",
"github_login": "khushidahi",
"run_id": "76b7ab1a-d3df-4daa-bc1e-25ae02f1d922",
"tier": "easy",
"dataset_id": "e5",
"status": "succ... | {
"score": {
"mean_loss": 2.5378607415428576,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.007083333383003871
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.53495584902742,
"example_count": 600,... |
3cbc736d-2a9d-4e67-be92-ec62454c1221 | easy | himalalps | 2026-08-06 09:14:45.230782+00:00 | succeeded | faa8aebcc50246446a58d92c8a81588a1b62a5d1138a56c362e79f3420f386d8 | 36,653 | null | """Algorithmically constrained latent recurrence for One Layer Deeper.
One masked cross-attention pass creates two independent representations:
``state_0`` from X only and an immutable context from N only. T is excluded
from both representations and is used solely to choose how many times the
learned recurrent MLP tr... | {
"id": "3cbc736d-2a9d-4e67-be92-ec62454c1221",
"created_at": "2026-08-06 09:14:45.230782+00:00",
"db_md5": "3eb99d52ff7311bfb8b17f0ca2bf0802",
"submitter": "Haoyu Tang",
"github_login": "himalalps",
"run_id": "25d4f1fe-b05a-495f-ab49-3bdf5dde272d",
"tier": "easy",
"dataset_id": "e2",
"status": "succe... | {
"score": {
"mean_loss": 3.017620325088501,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.012708333786576986
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 3.080491304397583,
"example_count": 300,... |
3cc238d6-82c8-404f-affd-1fd80e791078 | easy | amirmeisami | 2026-08-17 17:43:23.818438+00:00 | succeeded | 61c431e60c999d95a33f942cd68bb2a33eb69928ab00e162ec2862ea84d0c17a | 14,135 | null | """T021 Arm B: learned microprogram with a soft-routed operator population."""
from __future__ import annotations
import math
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import ModelSpec, OptimizerBundle, OptimizerSpec, Submission, TokenLossBatch, assert_model_state
D_MOD... | {
"id": "3cc238d6-82c8-404f-affd-1fd80e791078",
"created_at": "2026-08-17 17:43:23.818438+00:00",
"db_md5": "8fdf00067537a8536d48a5f992e07347",
"submitter": "Amir Meisami",
"github_login": "amirmeisami",
"run_id": "4d2b7d33-d8fd-46b8-a6ad-454c4a4d56ca",
"tier": "easy",
"dataset_id": "e7",
"status": "s... | {
"score": {
"mean_loss": 4.825305995819322,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.10597148028977627
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 5.065114526050847,
"example_count": 85,
... |
3cc983de-4e08-423d-8764-49fbacb7bfc5 | easy | DDanlov | 2026-08-14 08:02:01.198380+00:00 | succeeded | feed85fe7a9ee936e6981c8e66f85cadbcb40e2632d7fabc54df963cb1321f2e | 21,975 | null | import math
from typing import Optional
import torch
import torch.nn as nn
import torch.nn.functional as F
from torch.optim import Optimizer
try:
from benchmark import Submission, assert_model_state, OptimizerBundle
except ImportError:
try:
from client import Submission, assert_model_state, OptimizerBu... | {
"id": "3cc983de-4e08-423d-8764-49fbacb7bfc5",
"created_at": "2026-08-14 08:02:01.198380+00:00",
"db_md5": "65885f543ce44352632c3314ed7d0223",
"submitter": "DDanlov",
"github_login": "DDanlov",
"run_id": "ce545241-7b9b-468d-a5ad-75ec9ff76122",
"tier": "easy",
"dataset_id": "e5",
"status": "succeeded"... | {
"score": {
"mean_loss": 2.3770755068625324,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0141666666790843
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.376209303001652,
"example_count": 600,
... |
3cd37f1e-d36b-4068-80c0-1b0f18880971 | easy | shirvani-jr | 2026-08-27 00:17:44.584715+00:00 | succeeded | 92c3d18117c35e4e7e044bfd719dd8e8cc561a45ba0ce45433dda324ff60dd04 | 26,997 | null | """Adaptive learned discrete recurrent fabric for One Layer Deeper.
A gradient-trained, operation-free recurrent computational machine. The
model learns, from the competition's endpoint supervision: how to represent
the prompt, what to store in a bank of categorical registers, a reusable
transition circuit applied se... | {
"id": "3cd37f1e-d36b-4068-80c0-1b0f18880971",
"created_at": "2026-08-27 00:17:44.584715+00:00",
"db_md5": "60c2b33e32192769c3b628a2782d17db",
"submitter": "Ali",
"github_login": "shirvani-jr",
"run_id": "66eebfdf-192e-4a06-b67b-e2aad059648c",
"tier": "easy",
"dataset_id": "e1",
"status": "succeeded"... | {
"score": {
"mean_loss": 2.345192074775696,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.05166666582226753
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.3729653358459473,
"example_count": 100,... |
3cd79c5c-71b8-4d88-aecf-f349e969d856 | easy | DDanlov | 2026-08-11 19:31:23.669955+00:00 | failed | d7be3660ff277caac9ea5cb93904420d351d69703634edac70f4fa62209b16c3 | 16,175 | null | from __future__ import annotations
import sys
import os
import math
import time
import torch
import torch.nn as nn
import torch.nn.functional as F
from torch.optim.optimizer import Optimizer
VOCAB_SIZE = 17
DIGIT_OFFSET = 7
class RohanShampoo(Optimizer):
"""Self-contained RohanShampoo optimizer with eigenvalue ma... | {
"id": "3cd79c5c-71b8-4d88-aecf-f349e969d856",
"created_at": "2026-08-11 19:31:23.669955+00:00",
"db_md5": "666cd5dbe15cac6c2965ba7da2ad3e64",
"submitter": "DDanlov",
"github_login": "DDanlov",
"run_id": "bfc61ef7-ffb8-46ab-ad28-31a4a4f002cf",
"tier": "easy",
"dataset_id": "e1",
"status": "failed",
... | null
|
3cd9882c-7ad5-4ceb-b3cb-53115da41858 | easy | viridale | 2026-08-08 19:33:57.918877+00:00 | succeeded | 1847dbe6f59ac21d358f1ad0f1e68ed1db306904ef10304bc86d791bb57fc197 | 11,834 | null | """Factored learned spectral relation over a canonical global state.
hypothesis: v360's joint spectral relation is the intended new capability test, but
its materialized [source,destination,base,harmonic] activation truncated the CPU
ladder at T=1. The trigonometric angle-addition identity yields bit-equivale... | {
"id": "3cd9882c-7ad5-4ceb-b3cb-53115da41858",
"created_at": "2026-08-08 19:33:57.918877+00:00",
"db_md5": "77d6a686decca923013c9c60c25203df",
"submitter": "priormancer",
"github_login": "viridale",
"run_id": "594f830c-3cb7-4555-a0be-543f4481c204",
"tier": "easy",
"dataset_id": "e1",
"status": "succe... | {
"score": {
"mean_loss": 1.8000687956809998,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.10333333164453506
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 1.8321175575256348,
"example_count": 100... |
3ce7c9b4-135b-4482-916d-3b5af76b639a | easy | viridale | 2026-08-25 00:24:01.632739+00:00 | succeeded | 9b20b37fa807884a59698e09b268a4da279f206bbb40d4e8c97d9eb0e4a5695c | 18,749 | null | """Centered stochastic-Kaczmarz bilinear Transformer program.
hypothesis: v862's large-modulus failure is caused by forcing unsafe rank-completion
rows and expanding its unaliased chart to all sources; centered half-circle rows
supply scale-covariant rank-one Kaczmarz constraints without poisoned updates
axis:... | {
"id": "3ce7c9b4-135b-4482-916d-3b5af76b639a",
"created_at": "2026-08-25 00:24:01.632739+00:00",
"db_md5": "5bc2d14613e7a7119f6ca91a61dae614",
"submitter": "priormancer",
"github_login": "viridale",
"run_id": "60ad5158-6322-4eb3-bf01-b118775473d6",
"tier": "easy",
"dataset_id": "e6",
"status": "succe... | {
"score": {
"mean_loss": 0.0,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 1.0
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 0.0,
"example_count": 60,
"exact_accuracy": 1.0,
"c... |
3ce7f25b-7dba-4633-945a-f33c37eb9d5e | easy | DDanlov | 2026-08-24 21:08:46.499481+00:00 | succeeded | 6ba3caeca49b582fa0b48a2f8e3861cb81609b3848a8d8c9bdd9b10189c26865 | 46,194 | null | """
Unified Learned Query Slot Deliberation Model (91.2M Parameters) - Pure Skipless No-Norm Slot Transformer
Architecture:
- Deliberation Core: dim=1536, num_heads=24, d_ff=6144 (Pure Skipless SO(2)^{D/2} Givens Recurrence)
- Token Layout: Recurrent loop processes ONLY [Prompt (L_p), 4 Pause Registers] (Zero target to... | {
"id": "3ce7f25b-7dba-4633-945a-f33c37eb9d5e",
"created_at": "2026-08-24 21:08:46.499481+00:00",
"db_md5": "58d3d7e0b03d37a207822e9a2e3f86df",
"submitter": "DDanlov",
"github_login": "DDanlov",
"run_id": "ad40d245-f6d5-41de-b0dd-7a6542cc5592",
"tier": "easy",
"dataset_id": "e5",
"status": "succeeded"... | {
"score": {
"mean_loss": 3.020169966975269,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.006250000024835269
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 3.018768305201167,
"example_count": 600,... |
3ceb6b28-36f5-4a29-bfba-30a88e05ee73 | easy | jordanrubin | 2026-08-08 09:47:27.521592+00:00 | succeeded | 078cd8887c504bdcfcf580d442de465beb5b3a0ea5e1726a6c95fd2267731c91 | 16,615 | null | """Bidirectional transformer, D=256 x 3 blocks, grokking-tuned AdamW.
Design notes:
- Inputs are padded to config.max_seq_len inside forward and logits sliced
back, so every training batch presents one static shape (compile/cudagraph
friendly, and uniform kernel shapes even in eager).
- Attention mask gets an iden... | {
"id": "3ceb6b28-36f5-4a29-bfba-30a88e05ee73",
"created_at": "2026-08-08 09:47:27.521592+00:00",
"db_md5": "1f66d0c9d8b2addfecdef9a9aa694dae",
"submitter": "Jordan Rubin",
"github_login": "jordanrubin",
"run_id": "8e3e63c6-f328-40b7-8772-ac01e3919b93",
"tier": "easy",
"dataset_id": "e2",
"status": "s... | {
"score": {
"mean_loss": 3.8784420115603875,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.20749999474113187
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 6.264008244017137,
"example_count": 300,... |
3cf6fc70-9481-4f33-98ba-65e7c2abde9e | easy | rougee | 2026-08-22 06:49:20.753432+00:00 | succeeded | 6c9b07bcacb3145d3b8e055f801926549b3ed77e4bd888440a00a3132c05b027 | 5,126 | null | """Differentiable Recurrent Deep Looped Transformer for One Layer Deeper Benchmark."""
from __future__ import annotations
import math
from typing import Any
import torch
import torch.nn as nn
import torch.nn.functional as F
from torch import Tensor
from benchmark import (
ModelSpec,
OptimizerBundle,
Opti... | {
"id": "3cf6fc70-9481-4f33-98ba-65e7c2abde9e",
"created_at": "2026-08-22 06:49:20.753432+00:00",
"db_md5": "f467211de9229480ba0a1089ff6f7844",
"submitter": "Rougee",
"github_login": "rougee",
"run_id": "9038d301-75d5-4616-99fd-38e78e3466b6",
"tier": "easy",
"dataset_id": "e1",
"status": "succeeded",
... | {
"score": {
"mean_loss": 7.559138536453247,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.01666666753590107
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 9.495317459106445,
"example_count": 100,
... |
3d03cc0d-c802-4e46-a799-30be858a53c2 | easy | khushidahi | 2026-08-09 04:58:34.100715+00:00 | succeeded | 9438f731f22b2ddf6339fc7f8b2a31c737452aaa905dceb310b1b289b9732c06 | 23,846 | null | """R6 structured decimal-workspace model for One Layer Deeper.
The model does not implement multiplication, modular reduction, or the public
recurrence. It uses the public tokenizer structure to build learned, right-
aligned decimal tapes for N, X, and T, then applies a shared neural transition
to a mutable work tape.... | {
"id": "3d03cc0d-c802-4e46-a799-30be858a53c2",
"created_at": "2026-08-09 04:58:34.100715+00:00",
"db_md5": "55cdc6ddb28bc97b3d10167662242312",
"submitter": "khushidahi",
"github_login": "khushidahi",
"run_id": "d2dc85b0-afec-4619-b511-bfc014b56ff3",
"tier": "easy",
"dataset_id": "e5",
"status": "succ... | {
"score": {
"mean_loss": 2.1955673029428127,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.010000000024835268
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.1946492195129395,
"example_count": 60... |
3d064a65-b87c-44aa-9b0d-af3b1d2b9470 | easy | DDanlov | 2026-08-18 10:48:00.410161+00:00 | succeeded | 9e2e886a7f72218c6f6788250c79c4edcba3908cfb09a02021dbdad2941008ba | 23,584 | null | """
Official Submission for One Layer Deeper Challenge
Architecture: 4-Layer Recurrent Core Pure No-Norm No-Skip DKS Reasoning Model:
- 4 Distinct Pure No-Norm No-Skip Transformer Layers in the Recurrent Core: L1 -> L2 -> L3 -> L4
- Unrolled T times for serial modular squaring
- dim=512, 16 heads, d_ff=1024, RohanShamp... | {
"id": "3d064a65-b87c-44aa-9b0d-af3b1d2b9470",
"created_at": "2026-08-18 10:48:00.410161+00:00",
"db_md5": "d6174d2a9211082098f18de894eca987",
"submitter": "DDanlov",
"github_login": "DDanlov",
"run_id": "976f4834-20fc-4142-86c8-610d5902ea8a",
"tier": "easy",
"dataset_id": "e3",
"status": "succeeded"... | {
"score": {
"mean_loss": 2.3149252219914738,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.006875
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.3141200681870155,
"example_count": 800,
... |
3d068cc9-22fb-478d-adcb-dc7d9a20ba20 | easy | himalalps | 2026-08-10 01:33:08.808558+00:00 | succeeded | 025788022e440a086335df2b54772eefbdfa58d9068ef590b2c0396c142081bf | 84,552 | null | """Algorithmically constrained latent recurrence for One Layer Deeper.
One masked cross-attention pass creates two independent representations:
``state_0`` from X only and an immutable context from N only. T is excluded
from both representations and is used solely to choose how many times the
learned recurrent MLP tr... | {
"id": "3d068cc9-22fb-478d-adcb-dc7d9a20ba20",
"created_at": "2026-08-10 01:33:08.808558+00:00",
"db_md5": "f17ddf9da859bb64c60d9ff3ad80d2c6",
"submitter": "Haoyu Tang",
"github_login": "himalalps",
"run_id": "9058dc9a-21b5-4727-b931-9b300e381dc4",
"tier": "easy",
"dataset_id": "e5",
"status": "succe... | {
"score": {
"mean_loss": 2.179809294070478,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.006666666610787312
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.163473606109619,
"example_count": 600,... |
3d0de1ec-09a5-4377-a66b-8e8d5ce79b8d | easy | erdavis0 | 2026-08-29 00:59:55.924042+00:00 | succeeded | 516e9d9d041675bd73a4c56b612d0acbc7225da7b8d845cdabab52ec3efc5ebc | 20,130 | null | """Curvature table with a cached gentle modulus-relative residual."""
from __future__ import annotations
import math
import time
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
TokenLossBatch... | {
"id": "3d0de1ec-09a5-4377-a66b-8e8d5ce79b8d",
"created_at": "2026-08-29 00:59:55.924042+00:00",
"db_md5": "dac67f1d3596515bff6b89e826ff98e0",
"submitter": "Ethan",
"github_login": "erdavis0",
"run_id": "e631bbda-3de3-484d-979b-93e3bcc11993",
"tier": "easy",
"dataset_id": "e9",
"status": "succeeded",... | {
"score": {
"mean_loss": 0.10152619207474345,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.9766666889190674
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 4.249561970937066e-05,
"example_count": ... |
3d0e2426-0463-4eb0-9761-113779a0e62c | easy | TheYkk | 2026-08-04 11:42:45.122230+00:00 | succeeded | 7bf30e19d90e7f6cadb4ad67d0b2cd51634a4fd2d29a507be6f840a0dd78f8c1 | 3,628 | null | """Basic single-pass Transformer with PyTorch AdamW."""
from __future__ import annotations
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
assert_model_state,
)
D_MODEL = 128
NUM_HEADS = 4
... | {
"id": "3d0e2426-0463-4eb0-9761-113779a0e62c",
"created_at": "2026-08-04 11:42:45.122230+00:00",
"db_md5": "eaac69fe345f82973fcf392c504a40ff",
"submitter": "Kaan Karakaya",
"github_login": "TheYkk",
"run_id": "a2837e0f-11fa-4827-98f8-e7963d226209",
"tier": "easy",
"dataset_id": "e1",
"status": "succe... | {
"score": {
"mean_loss": 3.062563419342041,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.02333333296701312
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 4.053743362426758,
"example_count": 100,
... |
3d106ff0-a613-471e-80c8-c4214c396f39 | easy | ehonig | 2026-08-14 00:33:09.406291+00:00 | succeeded | a861c994826550eb9cb71193f14b147e5f0915f9a479fff32e5bfb5ab0addb70 | 46,921 | null | """DigitLoop — a place-value recurrent model for repeated modular squaring.
The task is ``y = x^(2^T) mod N`` with the answer read back tail aligned, one
decimal place value per token position from the right. A stack of attention and
MLP layers has no primitive for the two operations this needs — a partial-product
co... | {
"id": "3d106ff0-a613-471e-80c8-c4214c396f39",
"created_at": "2026-08-14 00:33:09.406291+00:00",
"db_md5": "483118408c645e6d3a1e944e31513465",
"submitter": "Edouardo Honig",
"github_login": "ehonig",
"run_id": "af15a951-f081-47fe-ac9f-3c02df727886",
"tier": "easy",
"dataset_id": "e1",
"status": "succ... | {
"score": {
"mean_loss": 3.9432953596115112,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.053333332762122154
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 4.239657878875732,
"example_count": 100... |
3d12b9a1-2ae5-4b71-92a0-edce2124bc2b | easy | G-AshwinKumar | 2026-08-14 00:26:04.833605+00:00 | succeeded | c8358ed6a0d6aa1fa23862b720aa7ca172e063c168187fd42fd0e828a1ef60d7 | 29,964 | null | """Recurrent squaring operator with a discrete residue state between iterations.
Continuous recurrent states drift: small errors in the learned dynamics compound
over iterations, so a model correct at its training depth degrades when run
longer. The measured fix is to make the state between iterations *discrete* --
re... | {
"id": "3d12b9a1-2ae5-4b71-92a0-edce2124bc2b",
"created_at": "2026-08-14 00:26:04.833605+00:00",
"db_md5": "50c8e631bd9d7eb3cb8e65a300ea8c58",
"submitter": "Ashwin Kumar",
"github_login": "G-AshwinKumar",
"run_id": "888a2d64-8456-495b-9487-703234b565d0",
"tier": "easy",
"dataset_id": "e3",
"status": ... | {
"score": {
"mean_loss": 2.2769249119669546,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.010625
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.2836082292596083,
"example_count": 800,
... |
3d137ae9-e61f-40d1-97b0-13d219c6872f | easy | jordanrubin | 2026-08-02 17:55:17.613463+00:00 | succeeded | 91144cb02675511de9da9d182d9d9f6735169b41ed65c57f484fc88cd15dc273 | 20,829 | null | """Looped bidirectional selective-SSM (Mamba-style), pure PyTorch.
The statefulness bet: input-dependent SSM dynamics along the token axis,
with the block weight-tied and applied NUM_LOOPS times in depth with input
injection — recurrence in depth is where the task's T-iteration lives.
The selective scan is computed i... | {
"id": "3d137ae9-e61f-40d1-97b0-13d219c6872f",
"created_at": "2026-08-02 17:55:17.613463+00:00",
"db_md5": "70450752edb21f2ce87fe48a126d1fc7",
"submitter": "Jordan Rubin",
"github_login": "jordanrubin",
"run_id": "f9a3dec2-8a5d-4f95-8364-7000e47e64d7",
"tier": "easy",
"dataset_id": "e1",
"status": "s... | {
"score": {
"mean_loss": 5.3395163545380635,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.02833333270003398
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 5.436426465191058,
"example_count": 100,... |
3d15709f-79b2-4ca7-98e9-e5c877ed3315 | easy | DDanlov | 2026-08-26 19:53:49.688156+00:00 | succeeded | 18d87f5fc4a5fc5771e695b7f870d41e05f444278793ffe4e5345947887efbd4 | 32,176 | null | import math
import time
from dataclasses import dataclass
from typing import Optional, Tuple, Dict, Any, List, Union
import torch
import torch.nn as nn
import torch.nn.functional as F
from torch.optim import Optimizer
try:
from onelayerdeeper.api import Submission, OptimizerBundle, ModelSpec, OptimizerSpec
except ... | {
"id": "3d15709f-79b2-4ca7-98e9-e5c877ed3315",
"created_at": "2026-08-26 19:53:49.688156+00:00",
"db_md5": "beb0fed2791a765c27629fec73c97c27",
"submitter": "DDanlov",
"github_login": "DDanlov",
"run_id": "7cff5f01-b42c-476c-9635-fd62f321a7e0",
"tier": "easy",
"dataset_id": "e5",
"status": "succeeded"... | {
"score": {
"mean_loss": 2.3091895593014895,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0141666666790843
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.3152721537602856,
"example_count": 600,... |
3d1672d5-cc50-4907-9cf8-ac96fbfcbd31 | easy | jordanrubin | 2026-08-08 10:59:52.080819+00:00 | succeeded | 19e434c7e9fc9d9d601292a0f5c0a4083813bc7790caf87c0582fc1bbcc2cad8 | 16,909 | 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": "3d1672d5-cc50-4907-9cf8-ac96fbfcbd31",
"created_at": "2026-08-08 10:59:52.080819+00:00",
"db_md5": "671c15c2db39c821999bae96a7defa51",
"submitter": "Jordan Rubin",
"github_login": "jordanrubin",
"run_id": "81058e42-7adc-43da-bf59-7b5f8575a791",
"tier": "easy",
"dataset_id": "e5",
"status": "s... | {
"score": {
"mean_loss": 4.810732796368626,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.004999999888241291
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 4.855567880008253,
"example_count": 600,... |
3d184ea7-9432-4b3b-bda1-f20cd760181d | easy | jordanrubin | 2026-08-10 20:34:02.324211+00:00 | succeeded | d66d3596d565e364e36b2e65e3e2d2e8d5bb226368b36c4f9dd3d714e86dfa54 | 19,332 | null | """Weight-tied MLP-cell loop over soft digit states (mlploop).
Cell = MLP over [expected digits, pairwise digit products, RNS residue
simplices (fixed differentiable mixing over Z_p), N digits]. The cell that
learns one-step modular squaring from direct pairs (52% unseen at 14k rows,
day-20 screen) inside the exact-T ... | {
"id": "3d184ea7-9432-4b3b-bda1-f20cd760181d",
"created_at": "2026-08-10 20:34:02.324211+00:00",
"db_md5": "881e1109fb87d3faf946c8d57e92a5d3",
"submitter": "Jordan Rubin",
"github_login": "jordanrubin",
"run_id": "c058412e-1a5b-4ee4-af14-818b812258de",
"tier": "easy",
"dataset_id": "e5",
"status": "s... | {
"score": {
"mean_loss": 3.373074622376227,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.006666666517655055
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.209696320674879,
"example_count": 600,... |
3d1a9cf8-ef87-4448-b031-f06e906f098d | easy | chad-atexpedient | 2026-08-04 21:06:19.178019+00:00 | succeeded | a4f05ac5d05b03b7cf50ebd2474e51b7f2a85664e61979120239079e980f757d | 3,877 | null | """R008C: one-pass, parameter-matched control for tied recurrence."""
from __future__ import annotations
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
assert_model_state,
)
WIDTH = 128
HEA... | {
"id": "3d1a9cf8-ef87-4448-b031-f06e906f098d",
"created_at": "2026-08-04 21:06:19.178019+00:00",
"db_md5": "5c169dab48359b59008af413caac05f0",
"submitter": "chad-atexpedient",
"github_login": "chad-atexpedient",
"run_id": "2dbf461d-6994-40d2-9bb5-613f5ea94e3f",
"tier": "easy",
"dataset_id": "e5",
"st... | {
"score": {
"mean_loss": 8.186189392457797,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.011250000049670536
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 8.487787875359368,
"example_count": 600,... |
3d1c2cc3-51c4-4346-b936-aff06a93a76e | easy | isaac0804 | 2026-08-06 16:22:10.230764+00:00 | succeeded | 08d463a8054a191b2c83f6e6249ad141c69aababad19ff8065e2a76f6c87cd0d | 6,923 | null | """Deep weight-tied recurrence (32 iterations) with per-iteration input
injection -- a genuinely deep test of the "scale recurring depth" hypothesis,
distinct from `e3_looped_weight_tied` (8 iterations, no injection).
Context: even E1 (a single FIXED N, zero cross-N interference) has topped
out around 8% across a full... | {
"id": "3d1c2cc3-51c4-4346-b936-aff06a93a76e",
"created_at": "2026-08-06 16:22:10.230764+00:00",
"db_md5": "66859a333fcbdf8ffa19b4203fc6f301",
"submitter": "Isaac Yong",
"github_login": "isaac0804",
"run_id": "bd19a8c3-5b47-4495-8711-a35b0c6b557f",
"tier": "easy",
"dataset_id": "e3",
"status": "succe... | {
"score": {
"mean_loss": 2.2706352317051115,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.008750000013969838
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.3139578304122264,
"example_count": 80... |
3d1cb341-36e0-4bdb-bd80-529c3c35b47a | easy | sapient-sapiens | 2026-08-19 17:57:33.989262+00:00 | failed | b5124ee43a4a584c14f6218e68a8a04a3a9ea7c9a655ae9c0167c27e969229fc | 16,199 | null | """Phase A baseline: legitimate adaptive digit-space modular squaring.
N, X, and T remain categorical token sequences. Marker locations define fields,
but no field is decoded into a number. A field-local encoder prevents T from
leaking into the shared arithmetic operator. The controller alone sees T.
Training uses on... | {
"id": "3d1cb341-36e0-4bdb-bd80-529c3c35b47a",
"created_at": "2026-08-19 17:57:33.989262+00:00",
"db_md5": "2ead600f2c7ef8d94afc075a9cb97edb",
"submitter": "Ertondy",
"github_login": "sapient-sapiens",
"run_id": "63458c23-f25f-43f7-8648-6e8d9d4fe519",
"tier": "easy",
"dataset_id": "e5",
"status": "fa... | null
|
3d1db8a2-22b8-4237-877c-bcda26d1b269 | easy | erdavis0 | 2026-08-29 00:35:54.075345+00:00 | succeeded | 4a4eae9f25e21ebdabf4bf8e5432668483e739c8ee6f1896443427ef060ef71e | 16,400 | null | """Sign-tied table with early full-window float32 curvature consistency."""
from __future__ import annotations
import math
import time
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
TokenLos... | {
"id": "3d1db8a2-22b8-4237-877c-bcda26d1b269",
"created_at": "2026-08-29 00:35:54.075345+00:00",
"db_md5": "860ac28bf0d44f71448d7e7162091339",
"submitter": "Ethan",
"github_login": "erdavis0",
"run_id": "ad88dc36-f2d5-4097-9d37-5636bc57a553",
"tier": "easy",
"dataset_id": "e2",
"status": "succeeded",... | {
"score": {
"mean_loss": 0.12206285482898238,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.9833333492279053
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 0.00010697615653043613,
"example_count":... |
3d3390ed-5e3e-45e4-9b15-1101af9dd301 | easy | khushidahi | 2026-08-05 22:22:56.255415+00:00 | succeeded | a4bd169d3948264b470b55a1a35ca90a5f62cab4b0dca00e775f2335e7351fbe | 25,703 | null | """R16 hard-transfer hybrid experiments for One Layer Deeper.
Motivated by complementary Hard behavior:
- the generic sequence recurrent model retained seen-N first-rung progress;
- the structured decimal-tape model retained OOD-N first-rung progress.
The dual variants co-train both learned representations, supervise... | {
"id": "3d3390ed-5e3e-45e4-9b15-1101af9dd301",
"created_at": "2026-08-05 22:22:56.255415+00:00",
"db_md5": "c5ff3b96f8bc822bf711b22b3a5cd682",
"submitter": "khushidahi",
"github_login": "khushidahi",
"run_id": "8b2909c0-235d-43ba-b3e4-b2c8f0a9a613",
"tier": "easy",
"dataset_id": "e5",
"status": "succ... | {
"score": {
"mean_loss": 2.1763937167890437,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.01250000017384688
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.177704334259033,
"example_count": 600,... |
3d3846b1-ca2c-4387-9a04-8c78921cd76c | easy | dlyr3 | 2026-08-23 19:51:50.797780+00:00 | failed | f94b635f6dd2b70c17e8a3d599ceaea226881d1a6201c012187cb986cfacb760 | 7,386 | null | """The M-layer of arXiv:2008.03936 applied to the repeated-squaring task.
M = B + T(U x) is built from the pooled prompt, exp(M) is taken with
torch.matrix_exp, and V + S exp(M) is read back out. The paper's activity regularisation is deliberately NOT applied: measured on e1,
any nonzero lambda drives ||exp(M)||_F... | {
"id": "3d3846b1-ca2c-4387-9a04-8c78921cd76c",
"created_at": "2026-08-23 19:51:50.797780+00:00",
"db_md5": "3f569f9edd87fb3629c5fd5274b314cf",
"submitter": "dlyr3",
"github_login": "dlyr3",
"run_id": "50df58cc-495f-476c-82d1-efeb05554c42",
"tier": "easy",
"dataset_id": "e1",
"status": "failed",
"st... | null
|
3d399107-31ac-497c-b438-7c0e482608e2 | easy | oupadhyay | 2026-08-15 21:45:18.816671+00:00 | succeeded | 7685f8865d8b205b1b0bdb24b3ee7e77bcf9f3b47438d3a399b5bc622cfb56cf | 3,988 | null | """Conditional optimizer screen: lr5."""
from __future__ import annotations
import torch
import torch.nn.functional as F
from torch import nn
from benchmark import ModelSpec,OptimizerSpec,OptimizerBundle,Submission,assert_model_state
W,D,H=4,64,32
class C:
def __init__(s,vocab_size,max_seq_len):s.vocab_size,s.... | {
"id": "3d399107-31ac-497c-b438-7c0e482608e2",
"created_at": "2026-08-15 21:45:18.816671+00:00",
"db_md5": "15672fe33fcc64bebb17b51e7067abc1",
"submitter": "Ojasw Upadhyay",
"github_login": "oupadhyay",
"run_id": "e5e18278-4fbf-49cb-b282-f5856faa4914",
"tier": "easy",
"dataset_id": "e5",
"status": "s... | {
"score": {
"mean_loss": 2.1933842478701138,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.005833333345750968
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.1597959043733743,
"example_count": 60... |
3d3e11a6-6779-4b12-93a4-880ce4cb442a | easy | khushidahi | 2026-08-20 02:40:43.895733+00:00 | succeeded | a2de1b2f88822c47bef94ed5ef99245630dede9fe906cd98b6ea57c293bae4f0 | 24,488 | null | """R285: parallel quadratic tied-codebook functional root.
This candidate is deliberately diagnostic. It trains on evaluator-owned final
labels at several depths so that one shared transition must compose. The model
never converts N or X to native numbers and never implements multiplication or
modular reduction.
Th... | {
"id": "3d3e11a6-6779-4b12-93a4-880ce4cb442a",
"created_at": "2026-08-20 02:40:43.895733+00:00",
"db_md5": "203f2a264bf17e647f69bfe1fcfe1e82",
"submitter": "khushidahi",
"github_login": "khushidahi",
"run_id": "aa44e58c-9812-4356-9734-fd8d155fd68a",
"tier": "easy",
"dataset_id": "e3",
"status": "succ... | {
"score": {
"mean_loss": 2.851355031842145,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.006875000013969838
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.819667881057701,
"example_count": 800,... |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.