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 |
|---|---|---|---|---|---|---|---|---|---|---|
26ea86fd-f5e2-45f8-834a-8ac9beec818e | easy | velocizapkar | 2026-08-16 06:38:53.400667+00:00 | succeeded | 38aa1c34a7a907338a11b720a8b0d1f9aad49d795bc4575cbdc9c373b6468d8a | 17,130 | null | """Geometry-calibrated, completely skipless Transformer with SOAP.
This is the strict L8 CutSkip+SOAP control with exactly two deep-signal
calibrations: the mimetic attention target uses beta=2, and GELU is centered
and variance-normalized for a standard Gaussian input. Every block remains
the replacement composition... | {
"id": "26ea86fd-f5e2-45f8-834a-8ac9beec818e",
"created_at": "2026-08-16 06:38:53.400667+00:00",
"db_md5": "892c567f96121f4e120a7a99355b06dc",
"submitter": "Aakanksh Zarapkar",
"github_login": "velocizapkar",
"run_id": "78971e40-d82c-4a17-8ca1-9dc1ad7e4486",
"tier": "easy",
"dataset_id": "e6",
"statu... | {
"score": {
"mean_loss": 5.657830238342285,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.03536036144942045
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 5.415378093719482,
"example_count": 60,
... |
26f28e6e-2dd3-430f-a51f-f50d11717751 | easy | yunjiangster | 2026-08-27 21:30:42.767900+00:00 | succeeded | 33e6bd17768403279623c1bfa039a90fff15fc414a56c22bd5532f7d290138ef | 14,270 | 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": "26f28e6e-2dd3-430f-a51f-f50d11717751",
"created_at": "2026-08-27 21:30:42.767900+00:00",
"db_md5": "5f218cf4042ba706d5c99172b93552e9",
"submitter": "Yunjiang Jiang",
"github_login": "yunjiangster",
"run_id": "039d3ab2-a614-4fdf-b0d8-c99cd0c29b0a",
"tier": "easy",
"dataset_id": "e3",
"status":... | {
"score": {
"mean_loss": 28.819966316223145,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.015624999534338713
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 28.705615997314453,
"example_count": 80... |
26f32f3d-e18a-4226-a39c-a41e6913f0b1 | easy | jayyala | 2026-08-16 23:04:06.167294+00:00 | succeeded | 9fb85c86563736265af0d6681567d86fe953dabca7a4c104c60db37f4450ec4e | 14,657 | null | """A scalable recurrent machine with canonical categorical byte states.
Each integer is represented by a short little-endian tape of 256-way slots.
One tied bidirectional mixing cell learns the unknown state transition from the
current tape and the prompt context, then its categorical output is fed back
for exactly T ... | {
"id": "26f32f3d-e18a-4226-a39c-a41e6913f0b1",
"created_at": "2026-08-16 23:04:06.167294+00:00",
"db_md5": "abe23542b6010fc06d199e62298e64a8",
"submitter": "jayyala",
"github_login": "jayyala",
"run_id": "b40d33b4-e767-44fe-9a7d-bfe246be79be",
"tier": "easy",
"dataset_id": "e3",
"status": "succeeded"... | {
"score": {
"mean_loss": 35.464773178100586,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.006874999729916453
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 35.321834564208984,
"example_count": 80... |
27045448-db34-46a6-809e-37eb7f2f5252 | easy | yashkant | 2026-08-09 04:35:08.680908+00:00 | succeeded | 77e9e9a00c94235a4475d7293bd0913658e4c1b021a6dd48373ac84494f79cb6 | 15,841 | null | """Round 236 RIMA source-literal z6/y1 schedule crossover smooth-inertia x/y/z candidate.
The only external scientific equation used here is the RIMA update from
arXiv:2603.05234: candidate=f(state, peer, x),
alpha=sigmoid(Linear(candidate)), and
state'=alpha*candidate+(1-alpha)*state. This is an independent implemen... | {
"id": "27045448-db34-46a6-809e-37eb7f2f5252",
"created_at": "2026-08-09 04:35:08.680908+00:00",
"db_md5": "9508f9001506e85c9df8d02593fa1ce6",
"submitter": "Yash Kant",
"github_login": "yashkant",
"run_id": "eae29c1a-2a96-49bf-9a2c-966468d673a7",
"tier": "easy",
"dataset_id": "e2",
"status": "succeed... | {
"score": {
"mean_loss": 3.2177268266677856,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.06645833514630795
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 3.8348779678344727,
"example_count": 300... |
270a319b-1c60-4426-898f-32fa3f72145e | easy | KaustubhKumar05 | 2026-08-25 15:37:51.631510+00:00 | succeeded | 0dd780da0f2643e57765b3d6e0e2a43ad3d3d1878ba500bfc9e4abda1f9875d1 | 17,167 | null | """conv_abacus: final_v1 + the CORRECT Abacus digit embedding.
final_v1 indexed positions from the end of the whole padded sequence, so the
units digit of N and the units digit of x landed on DIFFERENT embeddings --
the exact failure McLeish et al. (arXiv:2405.17399) identify. This uses a
table indexed by significance... | {
"id": "270a319b-1c60-4426-898f-32fa3f72145e",
"created_at": "2026-08-25 15:37:51.631510+00:00",
"db_md5": "3009ed5dc9d1fec3670f0014eba79c15",
"submitter": "koz",
"github_login": "KaustubhKumar05",
"run_id": "fce35998-418e-4de9-b1ea-f7a9ed3d45db",
"tier": "easy",
"dataset_id": "e3",
"status": "succee... | {
"score": {
"mean_loss": 7.496892060750131,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.008749999999999999
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 8.148782100805878,
"example_count": 800,... |
27116de9-0dc7-453a-9f51-2e02f34247ad | easy | AFeinfield | 2026-08-29 03:39:22.934243+00:00 | succeeded | 4b4c56256223963be2ebc499c37bad937694f819be26985ce028b62da73903f1 | 16,020 | null | """204: full-enumeration drills (all 100 digit pairs per call, zero
sampling noise) + drill weight 3 — kill the per-run table-snap lottery.
202: 201 + table lr 0.02 + Polyak EMA-shadow eval (decay .995) —
capture the snapped state through the late oscillation.
201: 200 + table-group lr 0.03 (GrokFast stays off) — the... | {
"id": "27116de9-0dc7-453a-9f51-2e02f34247ad",
"created_at": "2026-08-29 03:39:22.934243+00:00",
"db_md5": "c5d2b5ff5ac2382be1f2538a66de9081",
"submitter": "AFeinfield",
"github_login": "AFeinfield",
"run_id": "8e04ea69-69e5-4b25-9a67-79b2bd230584",
"tier": "easy",
"dataset_id": "e1",
"status": "succ... | {
"score": {
"mean_loss": 2.7976164676468938,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.792188774293928,
"example_count": 100,
"exac... |
27184b38-c5f2-4f1d-b59d-53a38712ba9c | easy | ehonig | 2026-08-19 21:18:16.232346+00:00 | succeeded | ac5493f1371683c864396404c6dd08181bbb08f97f1d7f8598144df44000f64b | 19,543 | null | """A plain Transformer for repeated modular squaring, and nothing else yet.
This is a deliberate restart. The previous submission accumulated a field
parser, a carry scan, a learned reciprocal, periodic features, a digit
bottleneck, a depth selector and two cell types, most of them tuned around a
memorisation table th... | {
"id": "27184b38-c5f2-4f1d-b59d-53a38712ba9c",
"created_at": "2026-08-19 21:18:16.232346+00:00",
"db_md5": "7e42a93fcd7d6009f54678fccc70e9c8",
"submitter": "Edouardo Honig",
"github_login": "ehonig",
"run_id": "46a52bc6-35d4-48a4-b83c-827b34f33806",
"tier": "easy",
"dataset_id": "e1",
"status": "succ... | {
"score": {
"mean_loss": 1.8690066933631897,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.07499999925494194
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 1.8190202713012695,
"example_count": 100... |
271904bb-1f39-4fa3-a75a-df939468f600 | easy | paulpister | 2026-08-05 03:47:58.606876+00:00 | succeeded | ab5ea73f0ec185e1dcdd6339546131842d0c6d5cbe2694f0ea633a149bb1cb49 | 21,396 | null | """Loop-T tied recurrence v18: v16 + per-row 8-step BPTT window (replaces random restart).
Design:
- Structural features computed from input_ids (field segment, digit place
value from the little end, reverse position from sequence end). Sinusoidal
place/revpos features so longer OOD-N numbers extrapolate smoothly.... | {
"id": "271904bb-1f39-4fa3-a75a-df939468f600",
"created_at": "2026-08-05 03:47:58.606876+00:00",
"db_md5": "0ed82a60f9ef8f45075560b121f469cd",
"submitter": "paulv",
"github_login": "paulpister",
"run_id": "93e7d41f-70c0-43c6-b866-ece2834aeb33",
"tier": "easy",
"dataset_id": "e2",
"status": "succeeded... | {
"score": {
"mean_loss": 8.029719352722168,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.14750000461935997
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 11.713325500488281,
"example_count": 300,... |
271aecf6-2e4f-4efb-a2b6-6bfd31b89d57 | easy | karanganesan | 2026-08-04 02:14:33.123719+00:00 | succeeded | 54244dd2411a1dbabf6a1d1487389d362d51a5df65125c076a84693044eb5932 | 11,980 | 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": "271aecf6-2e4f-4efb-a2b6-6bfd31b89d57",
"created_at": "2026-08-04 02:14:33.123719+00:00",
"db_md5": "a32501531606782100a874c225db355f",
"submitter": "Karan Ganesan",
"github_login": "karanganesan",
"run_id": "1d098c90-d4b8-4ae5-b5b9-ce4fc36fb032",
"tier": "easy",
"dataset_id": "e1",
"status": ... | {
"score": {
"mean_loss": 3.101262927055359,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.06666666641831398
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.8347432613372803,
"example_count": 100,... |
2724707a-1d0b-4270-8a11-3f68f6fa8b1a | easy | chad-atexpedient | 2026-08-05 16:05:51.191647+00:00 | succeeded | a80bdda2730df172a96339ebdcd70767d79007a5665a0a83d7016a406c27dffe | 11,099 | null | """T-M: R028F with explicit pooled-T memory-choice conditioning."""
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,
)
LANES = 6
CHANNEL... | {
"id": "2724707a-1d0b-4270-8a11-3f68f6fa8b1a",
"created_at": "2026-08-05 16:05:51.191647+00:00",
"db_md5": "c047cec69cb199e43f4770708b8cee02",
"submitter": "chad-atexpedient",
"github_login": "chad-atexpedient",
"run_id": "d037d73c-9ea7-471c-9606-11922c5459e8",
"tier": "easy",
"dataset_id": "e5",
"st... | {
"score": {
"mean_loss": 2.212742311652965,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.008333333333333333
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.2616078147973715,
"example_count": 600... |
2724a62e-8fc5-48c0-a6d3-fd516acd5442 | easy | liam-gb | 2026-08-12 11:54:31.311806+00:00 | succeeded | fe04ff8a6ae765fa195abec8bfbd9cbc50c8ca38f37f8a9061e67982119e31e7 | 14,353 | null | """rns_pm_gather_ct_b46r: the extended bank with per-prime temperature.
Adjacent residue classes on prime p are separated by 1 - cos(2*pi/p),
which shrinks like 1/p^2 — so one shared softmax temperature resolves
small primes and leaves large ones indistinguishable. Each prime gets
its own learnable temperature, initia... | {
"id": "2724a62e-8fc5-48c0-a6d3-fd516acd5442",
"created_at": "2026-08-12 11:54:31.311806+00:00",
"db_md5": "d0b639094423c56519b1dc5193478d28",
"submitter": "liam-gb",
"github_login": "liam-gb",
"run_id": "4eb73770-8755-4ad6-b6da-2a1f54fef989",
"tier": "easy",
"dataset_id": "e1",
"status": "succeeded"... | {
"score": {
"mean_loss": 3.832595705986023,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.3283333517611027
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 6.524946212768555,
"example_count": 100,
... |
27394c3a-6273-4c4c-8aa7-36b376808388 | easy | sapient-sapiens | 2026-08-22 18:33:05.414279+00:00 | succeeded | f30529160c07a01990778db141266fc818433b51adfecfa6935cce3561bea98b | 19,798 | null | """Batch-64 eight-layer T-private Q+V TRM with HybridMuon and SAM for E5.
Research hypothesis: batch 64 improves H100 utilization and fixed-clock learning
relative to the inherited batch 8 recipe.
Primary experimental variable: training batch size, 8 versus 64, with model,
optimizer, LR, SAM, and loss held fixed. The... | {
"id": "27394c3a-6273-4c4c-8aa7-36b376808388",
"created_at": "2026-08-22 18:33:05.414279+00:00",
"db_md5": "d0e469fdac4b4e4263c1ae354548a8bf",
"submitter": "Ertondy",
"github_login": "sapient-sapiens",
"run_id": "b15e98be-503c-4e92-9cb3-1acd3aaf4a74",
"tier": "easy",
"dataset_id": "e5",
"status": "su... | {
"score": {
"mean_loss": 2.202233204742634,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.002916666666666667
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.208672594863738,
"example_count": 600,... |
2748aa0c-f8d3-42f2-8428-252e58b63206 | easy | yashkant | 2026-08-15 13:02:55.412753+00:00 | succeeded | e2494728ba6de1810b534646ddb968eeab5904099d8673c05951a2cf883fbbc9 | 35,314 | 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": "2748aa0c-f8d3-42f2-8428-252e58b63206",
"created_at": "2026-08-15 13:02:55.412753+00:00",
"db_md5": "39df511e10424a63a2bf7764afd6af54",
"submitter": "Yash Kant",
"github_login": "yashkant",
"run_id": "9d23ba2b-2a31-4a9c-af23-816b0d1fb873",
"tier": "easy",
"dataset_id": "e3",
"status": "succeed... | {
"score": {
"mean_loss": 2.1185228155024864,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.016250000204890966
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.1180431524845647,
"example_count": 80... |
275c24a9-af66-4af8-ab64-803085a25033 | easy | oupadhyay | 2026-08-16 17:40:54.091847+00:00 | failed | 80ab5d49c20b52d5619d41d4956ac240309b41a301cb431cfcc10e5c746f4ddc | 4,268 | null | """Dynamic conditional T1-weight screen: weight4."""
from __future__ import annotations
import torch
import torch.nn.functional as F
from torch import nn
from benchmark import ModelSpec,OptimizerSpec,OptimizerBundle,Submission,TokenLossBatch,assert_model_state
W,D,H=None,96,48
class C:
def __init__(s,vocab_siz... | {
"id": "275c24a9-af66-4af8-ab64-803085a25033",
"created_at": "2026-08-16 17:40:54.091847+00:00",
"db_md5": "168a39ea48bae5545b5c48b797f6213f",
"submitter": "Ojasw Upadhyay",
"github_login": "oupadhyay",
"run_id": "9e6bb7fe-1fa8-4a39-b2e2-380d0756183c",
"tier": "easy",
"dataset_id": "e5",
"status": "f... | null
|
275ebd77-534c-4591-a4ee-c82e62b9883d | easy | TheYkk | 2026-08-04 12:02:14.732609+00:00 | succeeded | c18414d96afd02599e8e0d6bc831e98856b6afe3b5edee68dba2f1dc547cdcbc | 12,669 | null | """Weight-tied looped transformer with extrapolatable adaptive depth.
Design rationale
----------------
The label is a composition ``S^T(x)`` of an unknown per-instance step map.
A fixed-depth stack must learn every composition length as a separate
function; a weight-tied loop can instead learn the single step map onc... | {
"id": "275ebd77-534c-4591-a4ee-c82e62b9883d",
"created_at": "2026-08-04 12:02:14.732609+00:00",
"db_md5": "ffcc0168f27118f251ed1382260534fa",
"submitter": "Kaan Karakaya",
"github_login": "TheYkk",
"run_id": "c79df7ba-fae3-4965-9ed9-934bf54c0497",
"tier": "easy",
"dataset_id": "e1",
"status": "succe... | {
"score": {
"mean_loss": 4.764978647232056,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.03333333320915699
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 5.163893222808838,
"example_count": 100,
... |
27624009-45df-4837-8d94-cf0fab156d9a | easy | chad-atexpedient | 2026-08-05 16:04:17.425066+00:00 | succeeded | dfcf2ae6774699da5e8a757271e908f26f8b9cc19037e84ba2e569511769dce6 | 11,163 | null | """T-P: R028F with explicit pooled-T proposal conditioning."""
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,
)
LANES = 6
CHANNELS = 2... | {
"id": "27624009-45df-4837-8d94-cf0fab156d9a",
"created_at": "2026-08-05 16:04:17.425066+00:00",
"db_md5": "3cde56cf8a60f1400b06f7c82256f988",
"submitter": "chad-atexpedient",
"github_login": "chad-atexpedient",
"run_id": "9099e5fd-f4b8-43c8-b67f-4fd6f23ab3c8",
"tier": "easy",
"dataset_id": "e5",
"st... | {
"score": {
"mean_loss": 2.2609507285763737,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.00416666670391957
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.3076492264666366,
"example_count": 600... |
276bf500-da76-4e18-a4d5-46975fb6eafc | easy | DESU-CLUB | 2026-08-30 18:53:54.767408+00:00 | succeeded | b76551433ca464f807900f58503352b456867d65a3323cbb35f0afb40c624055 | 13,947 | null | """Discrete-state loop over a place-value tape, trained with orthogonalize-only Muon.
The recurrent state carried between squarings is the *digit string*, not a
residual stream. Each outer iteration reads (N, x) off an LSB-first place-value
tape, emits digit logits for one squaring, hard-argmaxes them (straight-throu... | {
"id": "276bf500-da76-4e18-a4d5-46975fb6eafc",
"created_at": "2026-08-30 18:53:54.767408+00:00",
"db_md5": "e7ead3cbb3a1c810b23a7d0f03b90169",
"submitter": "Low Keng Hoong, Warren",
"github_login": "DESU-CLUB",
"run_id": "aa61fc07-116a-4413-addd-a0db0e72d9f5",
"tier": "easy",
"dataset_id": "e6",
"sta... | {
"score": {
"mean_loss": 1.0836742177401504,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.6806306540966034
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 0.7644683122634888,
"example_count": 60,
... |
2772cd63-17ba-48cd-8d2c-dc736eec16eb | easy | sapient-sapiens | 2026-08-18 13:29:53.614947+00:00 | succeeded | 3ff9fc5a0835b3363ca4ae57f78751bc4bdf6a86e39ef66bf5ab39ca3bbe5efe | 29,237 | null | """Depth-quantized recurrence with a mixture-over-depth objective.
The recurrent state is a bank of right-aligned digit slots. Every operator
application is followed by a soft quantization back onto the token simplex, and
the digit logits produced by that quantization are the answer logits at that
depth. There is no s... | {
"id": "2772cd63-17ba-48cd-8d2c-dc736eec16eb",
"created_at": "2026-08-18 13:29:53.614947+00:00",
"db_md5": "50638b218fe7656fd6b8e87e4cee6538",
"submitter": "Ertondy",
"github_login": "sapient-sapiens",
"run_id": "9787384f-cd0e-4235-89eb-e3bad097bf11",
"tier": "easy",
"dataset_id": "e6",
"status": "su... | {
"score": {
"mean_loss": 3.7747011184692383,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.08355856314301491
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 3.7859718799591064,
"example_count": 60,... |
27745529-7512-4f3b-a321-ba51a29c169c | easy | khushidahi | 2026-08-16 18:47:03.504950+00:00 | succeeded | 3a94ccd0603b3628f3eebcfbdf5a09fe3034ca7a54f58f2811327bf50ad0c5b8 | 22,153 | null | """R76: composition-trained Neural GPU with a well-defined state tape.
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 re... | {
"id": "27745529-7512-4f3b-a321-ba51a29c169c",
"created_at": "2026-08-16 18:47:03.504950+00:00",
"db_md5": "97492f9b654d0e596fc23d53a7db8728",
"submitter": "khushidahi",
"github_login": "khushidahi",
"run_id": "9eaa5fdd-6772-4d47-92f3-94ff5761e44f",
"tier": "easy",
"dataset_id": "e10",
"status": "suc... | {
"score": {
"mean_loss": 1.2649784684181213,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.6316767930984497
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 1.082384705543518,
"example_count": 125,
... |
27883b1e-dde8-4cb9-8914-d1dafecc8dae | easy | khushidahi | 2026-08-20 05:00:10.233728+00:00 | succeeded | a0647be3a5fd942643d36a14bcacad57817659ae8c672d727c8be038cc5ab7e7 | 36,128 | null | """R318: short-operand bootstrap for a continuous polynomial 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 reduct... | {
"id": "27883b1e-dde8-4cb9-8914-d1dafecc8dae",
"created_at": "2026-08-20 05:00:10.233728+00:00",
"db_md5": "2a948d4ef7713850a102ddd952894828",
"submitter": "khushidahi",
"github_login": "khushidahi",
"run_id": "3f95fe0a-0090-465b-af99-8f2ce6007204",
"tier": "easy",
"dataset_id": "e3",
"status": "succ... | {
"score": {
"mean_loss": 2.45894644449998,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.00375
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.457812249585265,
"example_count": 800,
"ex... |
278be985-1437-4ae5-8573-c00594029b8f | easy | k-penchev | 2026-08-22 22:37:21.862022+00:00 | succeeded | 6dc0b2b3bd2075bb8054cea5618685304ad26ecdb9f7b9b5640554e856d0d25d | 4,755 | null | """Width-96 tied recurrence with a learned scalar state-update gate."""
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 = 96
... | {
"id": "278be985-1437-4ae5-8573-c00594029b8f",
"created_at": "2026-08-22 22:37:21.862022+00:00",
"db_md5": "a4f97ae47afeac419214fde654632f94",
"submitter": "Kaloyan Penchev",
"github_login": "k-penchev",
"run_id": "8422fe52-92c1-4d80-96e7-4171687488b2",
"tier": "easy",
"dataset_id": "e1",
"status": "... | {
"score": {
"mean_loss": 2.9016209840774536,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.044999999925494194
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 3.6954517364501953,
"example_count": 10... |
279f12a2-ad76-4593-aa44-0e30ac7a0002 | easy | amirmeisami | 2026-08-18 22:30:35.625363+00:00 | succeeded | 2db6197581d79588e22a7c1eb8fb8c909a771c8cd0ddfd742e9916eb16b20a5f | 12,938 | null | """T024 V3: D with a 75% differentiable visible-state boundary."""
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_MODEL, POLY_WIDT... | {
"id": "279f12a2-ad76-4593-aa44-0e30ac7a0002",
"created_at": "2026-08-18 22:30:35.625363+00:00",
"db_md5": "1470d5123ea384eb618db31a5603d5ec",
"submitter": "Amir Meisami",
"github_login": "amirmeisami",
"run_id": "4d25579f-fe39-49ae-8651-831b74e72d9a",
"tier": "easy",
"dataset_id": "e3",
"status": "s... | {
"score": {
"mean_loss": 4.907499918887285,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.005625
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 4.928549763130249,
"example_count": 800,
"... |
279f878a-ccdb-478f-b1ff-232dcc526d94 | easy | bernettorlando | 2026-08-04 04:38:26.284337+00:00 | succeeded | acf9f941bd8348c024d6180c50b9e75609de641590410bcde5fae1ef6e76dee9 | 7,254 | null | """Recurrent-memory architecture for the One Layer Deeper benchmark.
The prompt is encoded once. A small bank of learned memory tokens then performs
shared, gated updates while repeatedly reading the frozen prompt representation.
The answer is decoded from the resulting memory state.
"""
from __future__ import annot... | {
"id": "279f878a-ccdb-478f-b1ff-232dcc526d94",
"created_at": "2026-08-04 04:38:26.284337+00:00",
"db_md5": "6c5151fe20616ba0863c1dd0c0d33bf9",
"submitter": "bernettorlando",
"github_login": "bernettorlando",
"run_id": "09dcde51-24a3-4ac8-9532-7f52dbdcac84",
"tier": "easy",
"dataset_id": "e1",
"status... | {
"score": {
"mean_loss": 2.1836522817611694,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0833333320915699
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.19085955619812,
"example_count": 100,
... |
27a1a079-6981-4581-a526-1a44f9e53ed8 | easy | rishivg | 2026-08-12 23:19:44.298916+00:00 | succeeded | bcf5ba7d5da29256c51025c0c780187e2319e73dfbaa3fdf75f087b39a46d00b | 161,805 | null | """Looped transformer with place-value coordinates, Muon, and a clock-paced LR.
Design notes, in rough order of expected value:
1. Weight-tied recurrence with input injection. The task is function
composition of depth ``T``, so one block applied ``LOOPS`` times is a much
better prior than a single feed-forw... | {
"id": "27a1a079-6981-4581-a526-1a44f9e53ed8",
"created_at": "2026-08-12 23:19:44.298916+00:00",
"db_md5": "c93cf551db770988c738e507e87d17d4",
"submitter": "Rishi Gottumukkala",
"github_login": "rishivg",
"run_id": "8e30c87c-96e5-4983-8b95-3ba658cee059",
"tier": "easy",
"dataset_id": "e5",
"status": ... | {
"score": {
"mean_loss": 15.629239082336426,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.009166666772216558
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 16.058040618896484,
"example_count": 60... |
27a71931-6039-4e34-86de-a93ad5dab4ed | easy | khushidahi | 2026-08-11 13:36:24.980863+00:00 | succeeded | d2d4e2b84075f5bff9259df70f5b8b964c104ad924efd951769fae8d91949412 | 24,963 | null | """R51 learned bilinear diagonal-routing model.
The architecture uses a learned low-rank pair feature for decimal-place slots
i and j, then a parameter-free scatter-add to slot i+j. The pair features,
carry propagation, modulus conditioning, recurrent update, and output decoder
are all learned end-to-end.
No digit mu... | {
"id": "27a71931-6039-4e34-86de-a93ad5dab4ed",
"created_at": "2026-08-11 13:36:24.980863+00:00",
"db_md5": "8f8324f7aaba526cedcc19d42807c001",
"submitter": "khushidahi",
"github_login": "khushidahi",
"run_id": "dd84b742-1b1b-4257-bb5b-412da3cd5c8a",
"tier": "easy",
"dataset_id": "e5",
"status": "succ... | {
"score": {
"mean_loss": 2.1614936377055662,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.007916666666666667
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.162803904358047,
"example_count": 600... |
27b07cb3-e77c-4524-acb3-66402411ff55 | easy | DDanlov | 2026-08-31 20:26:32.690933+00:00 | succeeded | e926c9ea581c90232706c29639eb7d3b8d154f993a2c43ea08f1a584f3c0e475 | 32,412 | null | """
Unified Continuous Hopfield Equilibrium Propagation Deliberation Engine
=======================================================================
Architecture:
- Deliberation: Continuous Hopfield Multi-Layer Energy System (L=4, D=768, d_ffn=2304, H=16)
- Integration: In-Place Strang-Split Conformal Symplectic Velocit... | {
"id": "27b07cb3-e77c-4524-acb3-66402411ff55",
"created_at": "2026-08-31 20:26:32.690933+00:00",
"db_md5": "656e60223ca73dbc7f71881f853ccacd",
"submitter": "DDanlov",
"github_login": "DDanlov",
"run_id": "a37f466c-4191-43ee-9e76-8dc00004ae06",
"tier": "easy",
"dataset_id": "e2",
"status": "succeeded"... | {
"score": {
"mean_loss": 2.6137286532205697,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.006458333333333333
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.6129933286903904,
"example_count": 30... |
27b3141f-3f2f-45c8-bbed-777758003d9a | easy | khushidahi | 2026-08-05 22:22:28.525376+00:00 | succeeded | b5ada0202341f9d931076e376ff8c8b75865f6fd79a18f973a0adcca7916adf7 | 15,033 | 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": "27b3141f-3f2f-45c8-bbed-777758003d9a",
"created_at": "2026-08-05 22:22:28.525376+00:00",
"db_md5": "8e1012caddf2686d2329065ee8e95f86",
"submitter": "khushidahi",
"github_login": "khushidahi",
"run_id": "7e906866-4f31-4c26-bae2-27a5335c83a0",
"tier": "easy",
"dataset_id": "e5",
"status": "succ... | {
"score": {
"mean_loss": 2.1984591484069824,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.007083333563059568
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.193880796432495,
"example_count": 600... |
27b67ca6-14a9-423a-829b-2d6fcb9debbb | easy | Dandandan | 2026-08-05 16:43:32.394829+00:00 | failed | 07cbb65b515cbb3b5ebdc53118767301958b5517fbe13f0daac91726aa0da965 | 33,839 | null | """End-to-end recursive function learner for One Layer Deeper.
A population of randomly initialized polynomial register networks is reused
exactly once per requested outer step. The evaluator trains every coefficient
and the evidence selector through its normal forward/backward/optimizer loop.
There is no recurrence-... | {
"id": "27b67ca6-14a9-423a-829b-2d6fcb9debbb",
"created_at": "2026-08-05 16:43:32.394829+00:00",
"db_md5": "311a9e68684429c8e8a0d8f0c34e6e7a",
"submitter": "Daniël Heres",
"github_login": "Dandandan",
"run_id": "f94f41de-9321-476b-947a-b0d89ed48d81",
"tier": "easy",
"dataset_id": "e5",
"status": "fai... | null
|
27b854bb-854c-4c20-9e41-9ec1eabfd6a1 | easy | viridale | 2026-08-28 10:32:18.184479+00:00 | succeeded | 60cf54054f22f1d50f021079cbcb2c62db02d3def422ab5e67264b0162225d9f | 32,962 | null | """Group-deduplicated family-specialized exact-evidence slab.
hypothesis: v971's evidence selector deliberately finds one dominant same-modulus
group, and its loss intersects that group's endpoints against only one representative
program posterior. Because the deterministic controller is a function of N alone,... | {
"id": "27b854bb-854c-4c20-9e41-9ec1eabfd6a1",
"created_at": "2026-08-28 10:32:18.184479+00:00",
"db_md5": "31386a7ab5db0fa1c6df344e58113575",
"submitter": "priormancer",
"github_login": "viridale",
"run_id": "4adbe7a3-3973-4322-be99-725efaefc5c5",
"tier": "easy",
"dataset_id": "e1",
"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": 100,
"exact_accuracy": 1.0,
"... |
27bdcffe-b3b9-4098-82ac-93de45cfbfa0 | easy | 0Chris5R | 2026-08-01 07:27:37.600811+00:00 | succeeded | 6ed406defcd12b123f841e5de2d1a1ef5daacba5de67540206ae88934a4d110f | 23,440 | null | """Shared-digit Neural GPU with late full-objective EMA evaluation.
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 sam... | {
"id": "27bdcffe-b3b9-4098-82ac-93de45cfbfa0",
"created_at": "2026-08-01 07:27:37.600811+00:00",
"db_md5": "7ed1fe11c74b4d825bede9ad0332c3ad",
"submitter": "Chris ",
"github_login": "0Chris5R",
"run_id": "d87bbad6-81b6-4dae-bddd-910a5f7e003c",
"tier": "easy",
"dataset_id": "e5",
"status": "succeeded"... | {
"score": {
"mean_loss": 2.7696358924190196,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.027916666679084302
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.635360022831391,
"example_count": 600... |
27bfab15-5870-4e0d-87bd-15423047a553 | easy | erdavis0 | 2026-08-22 11:57:24.050456+00:00 | succeeded | 9aa21cfd116990232dcd127ca3622994e9056aef336f9240b31d0736aea53bd0 | 8,287 | null | """Learned geometric prompt moments with multiplicative scale binding."""
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 = 72
... | {
"id": "27bfab15-5870-4e0d-87bd-15423047a553",
"created_at": "2026-08-22 11:57:24.050456+00:00",
"db_md5": "74f468822dce7e749383257b9e669998",
"submitter": "Ethan",
"github_login": "erdavis0",
"run_id": "f6d143b4-258e-4673-b983-e06e0e20d865",
"tier": "easy",
"dataset_id": "e10",
"status": "succeeded"... | {
"score": {
"mean_loss": 6.874415874481201,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.01705050515010953
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 6.5611395835876465,
"example_count": 125,... |
27c4538a-0de2-4f13-b768-d57c86af9042 | easy | amirmeisami | 2026-08-16 17:58:37.415726+00:00 | failed | 0de2bd5ad50191db3f173ce267b66d5fabac07382305f6cebf994456c33c1e94 | 11,890 | null | """T021 Arm C: convolutional GRU-style recurrent state machine."""
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_MODEL, INNER_STE... | {
"id": "27c4538a-0de2-4f13-b768-d57c86af9042",
"created_at": "2026-08-16 17:58:37.415726+00:00",
"db_md5": "847cfeecbca1e9a7299e62eadea9675f",
"submitter": "Amir Meisami",
"github_login": "amirmeisami",
"run_id": "1af95d4d-6ed8-4c1e-9cd6-7d3192147054",
"tier": "easy",
"dataset_id": "e7",
"status": "f... | null
|
27c490a8-f8b9-4142-8ad3-58cc47b7b443 | easy | gauravmishra | 2026-08-10 03:46:28.325930+00:00 | succeeded | 534d7373bfa3a4a6d4564326eea7ec437841a494e0f5bbdfa5adc5026a4f9f6f | 12,545 | null | """Stage-1 information-gain K4 representation candidate."""
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,
assert_model_state,
)
POSITI... | {
"id": "27c490a8-f8b9-4142-8ad3-58cc47b7b443",
"created_at": "2026-08-10 03:46:28.325930+00:00",
"db_md5": "2721d0863048f4daafd2d2de127e3c1b",
"submitter": "Gaurav Mishra",
"github_login": "gauravmishra",
"run_id": "05ca124f-c339-4b19-8cbc-f454b2e2209f",
"tier": "easy",
"dataset_id": "e3",
"status": ... | {
"score": {
"mean_loss": 2.452943954586275,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.005000000018626451
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.5534502784895907,
"example_count": 800... |
27ce08c6-fe10-4b0a-be8f-17c463b542d5 | easy | shan23chen | 2026-08-08 02:02:58.587994+00:00 | succeeded | e416a1558d446bcfa2ebc256564d9e2c86ca9d10e985f62c910175a0be5ba188 | 20,747 | 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": "27ce08c6-fe10-4b0a-be8f-17c463b542d5",
"created_at": "2026-08-08 02:02:58.587994+00:00",
"db_md5": "e09369d2f27e2cbd110598a88b008263",
"submitter": "Shan Chen",
"github_login": "shan23chen",
"run_id": "b0baaf41-32eb-4b5e-9895-8dc4cce3bfdf",
"tier": "easy",
"dataset_id": "e5",
"status": "succe... | {
"score": {
"mean_loss": 2.8497625589370728,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.011666666716337204
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.824201822280884,
"example_count": 600... |
27cf0d8d-56ba-4cb1-9d54-23cad84aee4a | easy | yashkant | 2026-08-19 01:17:21.706527+00:00 | succeeded | 862146a1e9acbe92c8744760279022cd13aebb4d9dfb3fb9bb96711be0e3c16d | 41,871 | null | """Round2223 two-independent-branch low-capacity ensemble.
This source keeps the complete Round1829 nonlinear branch architecture,
SAM/Compositional-Muon optimizer, EMA, loss, and evaluator contract. It
instantiates two independently initialized width-288 branches and always
returns their symmetric arithmetic-mean lo... | {
"id": "27cf0d8d-56ba-4cb1-9d54-23cad84aee4a",
"created_at": "2026-08-19 01:17:21.706527+00:00",
"db_md5": "277a40ae4c484e54a3b1e41decf19bf8",
"submitter": "Yash Kant",
"github_login": "yashkant",
"run_id": "251586e3-d47d-486e-bf9b-33d6cc9f4177",
"tier": "easy",
"dataset_id": "e1",
"status": "succeed... | {
"score": {
"mean_loss": 2.737160325050354,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.05166666768491268
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.671588897705078,
"example_count": 100,
... |
27d63c05-b649-4bc4-bf9e-87a7e3c69629 | easy | shirvani-jr | 2026-08-30 23:27:25.823304+00:00 | succeeded | 5e04560fff6d76c6350ed134dce34bbeb5ddf03992d210d9cdc6dda6ff09212c | 17,544 | 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,
)
PERIOD_MINIMUM = 11
PERIOD_MAXIMUM = 100
E... | {
"id": "27d63c05-b649-4bc4-bf9e-87a7e3c69629",
"created_at": "2026-08-30 23:27:25.823304+00:00",
"db_md5": "cfec548a5587d3a2942b9c10220452bc",
"submitter": "Ali",
"github_login": "shirvani-jr",
"run_id": "3adf5538-a9b0-4e74-b7a7-dd121faa1c2c",
"tier": "easy",
"dataset_id": "e4",
"status": "succeeded"... | {
"score": {
"mean_loss": 2.3120841420424236,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.028958333308498063
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.58781654786042,
"example_count": 1200... |
27d748fe-4847-432f-b968-1ba07cefc44b | easy | shreyash-chonkie | 2026-08-26 12:04:03.413669+00:00 | succeeded | 355b6ffeefee378f7f00218a824f6bab6c023e1fb1cbcf4f7531d552ecb17ba7 | 11,156 | null | """Closed register machine over four quantized complex phases."""
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,
)
STATE_... | {
"id": "27d748fe-4847-432f-b968-1ba07cefc44b",
"created_at": "2026-08-26 12:04:03.413669+00:00",
"db_md5": "7b8504f9f746f3b503b8074d69f1ce62",
"submitter": "Shreyash",
"github_login": "shreyash-chonkie",
"run_id": "185edb3f-9408-4a6a-8d1e-3803f8b432a0",
"tier": "easy",
"dataset_id": "e1",
"status": "... | {
"score": {
"mean_loss": 1.896205723285675,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.018333333544433117
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 1.8305320739746094,
"example_count": 100... |
27d8a41f-8f34-4e01-ae4a-9ca5895bb804 | easy | sapient-sapiens | 2026-08-21 21:25:03.209008+00:00 | failed | 90425b1aea6f5785a5e70c77717b369ce9c69d7f23f913650e88c9b2760f4099 | 34,189 | null | """Depth-quantized recurrence with a mixture-over-depth objective.
The recurrent state is a bank of right-aligned digit slots. Every operator
application is followed by a soft quantization back onto the token simplex, and
the digit logits produced by that quantization are the answer logits at that
depth. There is no s... | {
"id": "27d8a41f-8f34-4e01-ae4a-9ca5895bb804",
"created_at": "2026-08-21 21:25:03.209008+00:00",
"db_md5": "dcc08c84e1698aad2f7d02a3b82c5b11",
"submitter": "Ertondy",
"github_login": "sapient-sapiens",
"run_id": "35451a75-b73e-4c41-8e20-cc569fa8e140",
"tier": "easy",
"dataset_id": "e9",
"status": "fa... | null
|
27db0c29-ac0b-4e5c-aa67-2ff3889a22f0 | easy | DDanlov | 2026-08-18 00:10:02.328887+00:00 | succeeded | 115d46ec00ec18546afcf1ba31d74ac788944592739c4e1a211477fa256dc823 | 23,102 | 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": "27db0c29-ac0b-4e5c-aa67-2ff3889a22f0",
"created_at": "2026-08-18 00:10:02.328887+00:00",
"db_md5": "ef8f2b4e45393e17f33e8d6b42c30e5d",
"submitter": "DDanlov",
"github_login": "DDanlov",
"run_id": "ef5b03d5-ab61-40ae-a60f-f7e300152fb7",
"tier": "easy",
"dataset_id": "e1",
"status": "succeeded"... | {
"score": {
"mean_loss": 1.9718461803653495,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.03500000017384688
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 1.9270480880096776,
"example_count": 100... |
27e4ab58-ccc5-46d7-b4b2-ef77182bb728 | easy | shirvani-jr | 2026-08-30 20:00:22.678217+00:00 | succeeded | 9d361d90d88d39b7f29834a63815ca1a132644c78825b83597751cd21eff5a23 | 22,456 | null | from __future__ import annotations
import math
import time
from contextlib import nullcontext
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import ModelSpec, OptimizerBundle, OptimizerSpec, Submission, TokenLossBatch, assert_model_state
MODEL_WIDTH = 128
ATTENTION_HEADS =... | {
"id": "27e4ab58-ccc5-46d7-b4b2-ef77182bb728",
"created_at": "2026-08-30 20:00:22.678217+00:00",
"db_md5": "25d51e49ad1f8ee04996f2c5a1e56c63",
"submitter": "Ali",
"github_login": "shirvani-jr",
"run_id": "e6445ff0-1e95-4f8e-990b-7edb9e4ba936",
"tier": "easy",
"dataset_id": "e3",
"status": "succeeded"... | {
"score": {
"mean_loss": 9.362115383148193,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.005624999757856131
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 9.229010581970215,
"example_count": 800,... |
27ebc240-b510-40e6-8199-ab739ce2fff9 | easy | jordanrubin | 2026-08-08 10:02:30.885940+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": "27ebc240-b510-40e6-8199-ab739ce2fff9",
"created_at": "2026-08-08 10:02:30.885940+00:00",
"db_md5": "1f66d0c9d8b2addfecdef9a9aa694dae",
"submitter": "Jordan Rubin",
"github_login": "jordanrubin",
"run_id": "2a5570eb-3ed2-4280-afba-6d537bb5f9c4",
"tier": "easy",
"dataset_id": "e5",
"status": "s... | {
"score": {
"mean_loss": 2.169141285035205,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.00458333323088785
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.1729772622809818,
"example_count": 600,... |
27f07bbd-ee00-4f83-ba04-4f3c50c7f9b1 | easy | EyimofeA | 2026-08-15 02:55:32.691375+00:00 | succeeded | cbf327b36f8306a10ccfcb52bef481c1ee2d7b22bf1de93013ab8fb818a88268 | 13,707 | null | """Final-label-only Neural GPU square phase plus learned reduction phase.
The model uses generic learned scratch lanes and tied local updates. It has no
arithmetic trace, carry target, task solver, or hard-coded numeric transition.
Training uses only evaluator-provided final labels.
"""
from __future__ import annotati... | {
"id": "27f07bbd-ee00-4f83-ba04-4f3c50c7f9b1",
"created_at": "2026-08-15 02:55:32.691375+00:00",
"db_md5": "0e073167a13091b2d07cd3317bef757a",
"submitter": "mof",
"github_login": "EyimofeA",
"run_id": "e600a313-0a75-4a19-a1e0-af1ca8b8161b",
"tier": "easy",
"dataset_id": "e5",
"status": "succeeded",
... | {
"score": {
"mean_loss": 3.203389813386896,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.005416666666666667
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 3.1435380644862425,
"example_count": 600... |
27f6e3e1-32c1-41a6-9e1a-b44fe58a8cd8 | easy | DDanlov | 2026-08-26 20:29:51.189512+00:00 | succeeded | fb853c037373e0cd33b529521148702319bc36f6b7b5705ff5491f7a6f73c6aa | 34,343 | 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": "27f6e3e1-32c1-41a6-9e1a-b44fe58a8cd8",
"created_at": "2026-08-26 20:29:51.189512+00:00",
"db_md5": "df9b709ec5f94285fe7d476a18f5c971",
"submitter": "DDanlov",
"github_login": "DDanlov",
"run_id": "68163595-612b-41e0-87fa-5eebf22c2bd8",
"tier": "easy",
"dataset_id": "e5",
"status": "succeeded"... | {
"score": {
"mean_loss": 2.2928517975338427,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0020833333457509675
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.289445890974036,
"example_count": 60... |
27fca607-aa91-47d4-95c0-7bd993af5f6b | easy | karanganesan | 2026-08-08 04:00:21.832716+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": "27fca607-aa91-47d4-95c0-7bd993af5f6b",
"created_at": "2026-08-08 04:00:21.832716+00:00",
"db_md5": "3e8439ba8ab561733e7dc883fab4ecb2",
"submitter": "Karan Ganesan",
"github_login": "karanganesan",
"run_id": "c265ad39-bb1a-4db5-88b8-3a360f8d71d5",
"tier": "easy",
"dataset_id": "e3",
"status": ... | {
"score": {
"mean_loss": 9.73068380355835,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.009374999674037099
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 9.625393867492676,
"example_count": 800,
... |
2801696b-14fb-4411-8e90-057d057ca146 | easy | karanganesan | 2026-08-05 12:09:34.696824+00:00 | succeeded | 646a054ea7deeddec66b19823252ac47c4f3f282f2324bdf5771af44145e8b73 | 22,063 | 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": "2801696b-14fb-4411-8e90-057d057ca146",
"created_at": "2026-08-05 12:09:34.696824+00:00",
"db_md5": "214184c3194912fec1e2b1b37cdc453c",
"submitter": "Karan Ganesan",
"github_login": "karanganesan",
"run_id": "50eb4620-22e0-4329-a3a1-c2f761a11165",
"tier": "easy",
"dataset_id": "e1",
"status": ... | {
"score": {
"mean_loss": 8.605176210403442,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.02166666742414236
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 7.101634502410889,
"example_count": 100,
... |
28036472-45f7-46c9-816b-5f80020307e6 | easy | heathsanchez | 2026-08-17 07:16:30.249923+00:00 | succeeded | e05a27259a98cdd802bc2ae3b0802bcdcc6a54a295247c2e3fde776c73e65e44 | 6,224 | null | """E0005: explicit mutable workspace state, anchored to immutable context.
Train: one cheap workspace update.
Eval: four tied workspace updates.
The token stream is encoded once; mutable workspace is a separate latent state.
"""
from __future__ import annotations
import torch
import torch.nn.functional as F
from torch... | {
"id": "28036472-45f7-46c9-816b-5f80020307e6",
"created_at": "2026-08-17 07:16:30.249923+00:00",
"db_md5": "c80b0c0d6030a1262bcf43029b2f0787",
"submitter": "Heath Sanchez",
"github_login": "heathsanchez",
"run_id": "c3fdae8c-f49b-4f67-a4b5-5f2c10d612c5",
"tier": "easy",
"dataset_id": "e1",
"status": ... | {
"score": {
"mean_loss": 4.450322866439819,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0833333320915699
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 4.501961708068848,
"example_count": 100,
... |
2809f0f3-1f09-413f-bf94-60259d5ab71f | easy | DDanlov | 2026-08-17 11:21:31.799793+00:00 | succeeded | c6a5a0ed98c31b15829824ca7be07993f2f44cb72ad8b78c5111541b5e623f7d | 28,812 | null | """
Scalable Recurrent Deliberation Transformer (ScalableDEPHNet)
============================================================
Architecture:
- Pure No-Norm: Single unit RMS normalization strictly isolated to the input token embedding layer.
Zero LayerNorm / RMSNorm inside recurrent deliberation blocks.
- Pure No-Skip... | {
"id": "2809f0f3-1f09-413f-bf94-60259d5ab71f",
"created_at": "2026-08-17 11:21:31.799793+00:00",
"db_md5": "7fb69c6d979828323fb21d9082c54de6",
"submitter": "DDanlov",
"github_login": "DDanlov",
"run_id": "6bf98ede-e729-4ef0-8634-59164f813f81",
"tier": "easy",
"dataset_id": "e1",
"status": "succeeded"... | {
"score": {
"mean_loss": 2.093591469481699,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.07999999741713207
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.0731871128082275,
"example_count": 100,... |
280a1d09-dbfb-40a9-afbe-6311067bf2e1 | easy | lpbb | 2026-08-14 10:23:04.192289+00:00 | succeeded | 70b3b22416fbf016d9759c4bea52a11476f9d9863703b1dc344d44bfd73c218e | 8,059 | null | """CHEAT L1 — the model receives x^(2^(T-1)) mod N (the state one squaring before
the answer) as extra hint tokens computed in forward. It must LEARN the remaining
single squaring-mod step. Score isolates: is one step learnable in 60s at all?"""
from __future__ import annotations
import math
import torch
import torc... | {
"id": "280a1d09-dbfb-40a9-afbe-6311067bf2e1",
"created_at": "2026-08-14 10:23:04.192289+00:00",
"db_md5": "b7e04eaaefd3e612b1c382ae971c9147",
"submitter": "Lpbb",
"github_login": "lpbb",
"run_id": "6fe5bdb1-400c-4818-8ab6-7e6826a2178d",
"tier": "easy",
"dataset_id": "e3",
"status": "succeeded",
"s... | {
"score": {
"mean_loss": 5.876541937558356,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.07187500001862646
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 6.275763867530293,
"example_count": 800,
... |
2818e1a2-c204-4413-bf3e-a2dc12784c11 | easy | viridale | 2026-08-05 00:34:47.592294+00:00 | succeeded | 165c0697752c06ec637c8c045554be1f9d45bdcfe1988cfe77df993133429ddd | 10,336 | null | """Clean continuous two-dimensional scratch-grid recurrent machine.
hypothesis: transferable serial computation needs a generic work dimension and learned
local contraction/error correction; dense recurrence, one-dimensional ConvGRU, and
hard token quantization each lacked that complete substrate.
axis: arch
t... | {
"id": "2818e1a2-c204-4413-bf3e-a2dc12784c11",
"created_at": "2026-08-05 00:34:47.592294+00:00",
"db_md5": "a71a407abcf02609c034e54afd2c14bf",
"submitter": "priormancer",
"github_login": "viridale",
"run_id": "e9e49c63-f064-4823-a49f-b05a38a261f0",
"tier": "easy",
"dataset_id": "e1",
"status": "succe... | {
"score": {
"mean_loss": 5.048031806945801,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.021666666492819786
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 5.1671552658081055,
"example_count": 100... |
2831ec45-f4b1-44c8-bcdc-9940344681f5 | easy | jordanrubin | 2026-08-11 01:35:44.991659+00:00 | succeeded | aa0c446a01b1c50a3739ac906b125fbe05d6b1cd9cf0a23d1a564e062285e8d6 | 21,670 | 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": "2831ec45-f4b1-44c8-bcdc-9940344681f5",
"created_at": "2026-08-11 01:35:44.991659+00:00",
"db_md5": "344e901468d7fc4aff5006fca563aa6e",
"submitter": "Jordan Rubin",
"github_login": "jordanrubin",
"run_id": "e8f6110f-0d5c-4122-ae17-1d5d0918a99e",
"tier": "easy",
"dataset_id": "e5",
"status": "s... | {
"score": {
"mean_loss": 2.552079997017943,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.006249999860301614
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.154432564572903,
"example_count": 600,... |
28332fa5-03f9-4d90-b64f-2ab8050075c0 | easy | k-penchev | 2026-08-22 22:32:13.896857+00:00 | succeeded | e1a6f936814636c74a3f8cd4675e59d4de79cb6be0400b044fa4499ac119fab0 | 4,630 | null | """Width-96 model trained at two loops and evaluated at four."""
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 = 96
NUM_HEA... | {
"id": "28332fa5-03f9-4d90-b64f-2ab8050075c0",
"created_at": "2026-08-22 22:32:13.896857+00:00",
"db_md5": "81cc12f29f22e1c8acf5ee1155ac2fdb",
"submitter": "Kaloyan Penchev",
"github_login": "k-penchev",
"run_id": "1e71b9c0-89ec-4e72-b731-3cb84af0cac5",
"tier": "easy",
"dataset_id": "e1",
"status": "... | {
"score": {
"mean_loss": 2.7441232204437256,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.05333333043381572
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.83447003364563,
"example_count": 100,
... |
283a281a-9540-44e4-a6df-18179b7f6b29 | easy | sapient-sapiens | 2026-08-25 07:55:53.306215+00:00 | succeeded | 6e0722b4a6d66f3bc76f52ba7fcc58b7b70dc7293ba0a412fc0bdda37f56ccf8 | 18,821 | null | """Standalone fixed-depth small-model E5 scaling candidate.
Generated variants differ only in the CONFIG literal. Inputs remain
categorical token/group/left-relative embeddings. No N/X/T field is decoded
numerically, fixed depth is independent of T, and only the final answer is a
training target.
"""
from __future__ ... | {
"id": "283a281a-9540-44e4-a6df-18179b7f6b29",
"created_at": "2026-08-25 07:55:53.306215+00:00",
"db_md5": "b6a4f19c2bd0a0dc58160d672311e67a",
"submitter": "Ertondy",
"github_login": "sapient-sapiens",
"run_id": "b31790f8-09b6-46b3-a26e-bbe31fe2372b",
"tier": "easy",
"dataset_id": "e5",
"status": "su... | {
"score": {
"mean_loss": 2.5968170223885325,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.006666666666666666
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.979924553178351,
"example_count": 600... |
283e5fe2-5e37-442d-8743-59f4c4054ff5 | easy | oupadhyay | 2026-08-12 06:27:34.707142+00:00 | succeeded | fd8409bbda6f83a6fbaad320745286821341d7a8cf6dd1ebde18baaa2694718e | 7,871 | null | """One-block axial spatial-gating categorical register for the Easy tier."""
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, N_MARK, X_MARK, T_MAR... | {
"id": "283e5fe2-5e37-442d-8743-59f4c4054ff5",
"created_at": "2026-08-12 06:27:34.707142+00:00",
"db_md5": "6d83bac7d942a02f5e30edd11ae36676",
"submitter": "Ojasw Upadhyay",
"github_login": "oupadhyay",
"run_id": "e441fd65-a365-440b-884f-8b91924b7c49",
"tier": "easy",
"dataset_id": "e5",
"status": "s... | {
"score": {
"mean_loss": 2.1974449159812792,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.008333333358168602
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.1494535170328457,
"example_count": 60... |
283ff1e0-96a1-435c-9c4c-1da22d085075 | easy | lpbb | 2026-08-14 10:28:02.430137+00:00 | succeeded | a785757ab3811644d72abaece215d9801542b0f67fea1173e0b195da128bb874 | 8,075 | null | """CHEAT L1c — L1 (penultimate state given, learn one squaring-mod step) with a
bilinear MLP: down(u * v). Tests whether the multiplicative primitive lets the
arithmetic circuit beat memorization in the same 60s."""
from __future__ import annotations
import math
import torch
import torch.nn.functional as F
from torc... | {
"id": "283ff1e0-96a1-435c-9c4c-1da22d085075",
"created_at": "2026-08-14 10:28:02.430137+00:00",
"db_md5": "07a48f01be6f25af34b8cfe2ef2cf51a",
"submitter": "Lpbb",
"github_login": "lpbb",
"run_id": "fd20654c-26af-4913-add3-a0ccce822cb0",
"tier": "easy",
"dataset_id": "e3",
"status": "succeeded",
"s... | {
"score": {
"mean_loss": 5.627695397185786,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0993750000372529
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 5.594822198140005,
"example_count": 800,
... |
2847cf90-b10a-44a1-96f5-4a9edea7e566 | easy | khushidahi | 2026-08-06 03:32:13.887376+00:00 | succeeded | c61f9c909cbe2d3bfd0867c3c3f5dda3393455d6cfb3771a31094a08b8705bc5 | 14,365 | 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": "2847cf90-b10a-44a1-96f5-4a9edea7e566",
"created_at": "2026-08-06 03:32:13.887376+00:00",
"db_md5": "44a81ee90264474bd8c72b1c2f205530",
"submitter": "khushidahi",
"github_login": "khushidahi",
"run_id": "e1b67870-851e-4124-b33f-78fe647f852e",
"tier": "easy",
"dataset_id": "e5",
"status": "succ... | {
"score": {
"mean_loss": 2.1798452138900757,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.00541666685603559
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.1764118671417236,
"example_count": 600... |
28497240-fd96-42d3-b55f-36d1aabed508 | easy | sapient-sapiens | 2026-08-14 17:08:24.976061+00:00 | succeeded | 28ba469aae3e42ca679270105f219b6bd47433fa80f83dc154b3c3e55e44c7c3 | 14,299 | null | """Digit-space recurrent operator for repeated modular squaring.
Nothing here converts a token into a decimal number. Digits stay as embeddings,
so what a digit is worth, and every carry and reduction, is learned. That also
sidesteps a hard numeric wall: at Hard's modulus scale `x^2` runs past the 53-bit
exact-integer... | {
"id": "28497240-fd96-42d3-b55f-36d1aabed508",
"created_at": "2026-08-14 17:08:24.976061+00:00",
"db_md5": "1efadda9e01ae84ec2857432a666c542",
"submitter": "Ertondy",
"github_login": "sapient-sapiens",
"run_id": "659be596-7869-43c4-b232-961a0a8f6bef",
"tier": "easy",
"dataset_id": "e5",
"status": "su... | {
"score": {
"mean_loss": 2.18554201327869,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.005416666691501935
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.1977818232748008,
"example_count": 600,... |
284bce30-58bf-48ed-8ff2-eba6b336dd46 | easy | lzy54 | 2026-08-29 08:38:34.898489+00:00 | succeeded | 2198c3e036cb63ce07ed516c32c16517b7cd5fafc3e2aa65b0b27ed0fb8ccf49 | 4,032 | null | """Phase 2 C1: four untied blocks with additive input recall."""
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_HE... | {
"id": "284bce30-58bf-48ed-8ff2-eba6b336dd46",
"created_at": "2026-08-29 08:38:34.898489+00:00",
"db_md5": "859baf35963cf1712b065b05e50b68ce",
"submitter": "lzy54",
"github_login": "lzy54",
"run_id": "d6e38feb-1fcc-4382-b3b1-4728e13dad31",
"tier": "easy",
"dataset_id": "e8",
"status": "succeeded",
... | {
"score": {
"mean_loss": 3.0626901388168335,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.03319964371621609
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 3.150096893310547,
"example_count": 85,
... |
284f1dda-0403-4001-963e-edb679bbed6c | easy | Maxwells-Demons | 2026-08-06 04:50:44.169642+00:00 | succeeded | cd0d7a0515977cd35439f670bed832ecc8668e9d818ebe677ba0957220275118 | 11,988 | null | """Value-bottleneck recurrent model with pinned digit head (v10 champion).
State is a distribution over 3 decimal digit slots; a shared MLP learns the
single-step squaring-mod transition. The output head is a fixed scatter of
digit logits into vocab slots so the recurrent code is forced to equal true
digits, and slots... | {
"id": "284f1dda-0403-4001-963e-edb679bbed6c",
"created_at": "2026-08-06 04:50:44.169642+00:00",
"db_md5": "9b95774d1bcd31d4d72012a38fcd8fb8",
"submitter": "Maxwells_Ayakashi",
"github_login": "Maxwells-Demons",
"run_id": "a4ab93e5-4f7c-43bc-a9b2-74654c1ccc58",
"tier": "easy",
"dataset_id": "e1",
"st... | {
"score": {
"mean_loss": 1.5731387436389923,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.871666669845581
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 0.19611364603042603,
"example_count": 100,... |
285427e5-67f8-4e09-a246-03e8bc4c084f | easy | newjordan | 2026-08-09 19:29:39.387778+00:00 | succeeded | 6fce754dce70db68d5c33a7be56f6dc86c2e1438e4a131f9e0f97b51ca4dd7cb | 94,070 | 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": "285427e5-67f8-4e09-a246-03e8bc4c084f",
"created_at": "2026-08-09 19:29:39.387778+00:00",
"db_md5": "12165f416be599d050cf42ef34310469",
"submitter": "Frosty40",
"github_login": "newjordan",
"run_id": "26157f03-5c1d-44ae-9e07-2862bd7e73d1",
"tier": "easy",
"dataset_id": "e5",
"status": "succeed... | {
"score": {
"mean_loss": 6.064125786216308e-07,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 1.0
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 6.053344656104736e-07,
"example_count": 600,
... |
285a19ef-51cf-46a1-b492-5a140183f013 | easy | shirvani-jr | 2026-08-22 13:26:55.497613+00:00 | succeeded | 0c6dbdc495254e0bce7220529f7d2fb3cede38c5c2805a4e8d7a5defe87a5b40 | 24,165 | null | """Looped exact-arithmetic digit machine.
The recurrent state is the decimal digit string of the current residue. One
loop application performs one modular squaring, so depth extrapolates in T by
construction. Every functional primitive — digit-by-digit product, full
adder, full subtractor, digit comparator, and com... | {
"id": "285a19ef-51cf-46a1-b492-5a140183f013",
"created_at": "2026-08-22 13:26:55.497613+00:00",
"db_md5": "484dcf2452d8db1e8c365a23fff90dcf",
"submitter": "Ali",
"github_login": "shirvani-jr",
"run_id": "80ac88db-4339-43e8-abf4-30811179e270",
"tier": "easy",
"dataset_id": "e6",
"status": "succeeded"... | {
"score": {
"mean_loss": 0.016209575347602367,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 1.0
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 0.016157737001776695,
"example_count": 60,
"... |
285d29e2-132e-4eb5-b323-0aa734e24e12 | easy | alirezashirvani-jr | 2026-08-30 00:10:37.083346+00:00 | succeeded | 672293a76323da62e296a3ee3f10c6ca38ac913874438ad7921545a2384944f0 | 31,682 | null | from __future__ import annotations
import math
import time
from contextlib import nullcontext
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
TokenLossBatch,
assert_model_state,
)
MODEL_... | {
"id": "285d29e2-132e-4eb5-b323-0aa734e24e12",
"created_at": "2026-08-30 00:10:37.083346+00:00",
"db_md5": "310e8a85d3554c9bd27bdd0854d43589",
"submitter": "alirezashirvani-jr",
"github_login": "alirezashirvani-jr",
"run_id": "d5f2c8b6-7b89-4ae1-a911-c64eba147efa",
"tier": "easy",
"dataset_id": "e1",
... | {
"score": {
"mean_loss": 0.9935503602027893,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.7599999904632568
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 0.6011372804641724,
"example_count": 100,... |
2863a6f7-c211-473f-8213-1b1b11c00667 | easy | KaustubhKumar05 | 2026-08-28 06:22:27.909013+00:00 | succeeded | a8477a9210054548490820eec5f05c5156e50f1e96664f08941b12b03bd49619 | 11,234 | null | """cell_op2 -- ONE shared squaring operator with internal depth, applied T times.
WHY THIS SHAPE (verified on the real data, not assumed):
answer(T=t) == x squared t times mod N 20000 / 20000 examples
answer(T=b) == answer(T=a)^(2^(b-a)) mod N 1269 / 1269 pairs
So the entire T = 1,2,4,8,...,64... | {
"id": "2863a6f7-c211-473f-8213-1b1b11c00667",
"created_at": "2026-08-28 06:22:27.909013+00:00",
"db_md5": "963dfe68bd1ba25e49273a353947ba54",
"submitter": "koz",
"github_login": "KaustubhKumar05",
"run_id": "53e31f8e-1724-44f8-9e49-26ec537eface",
"tier": "easy",
"dataset_id": "e4",
"status": "succee... | {
"score": {
"mean_loss": 3.88655697926328,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.004479166679084301
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 4.563136280503627,
"example_count": 1200,... |
286dfa92-fe8e-41c3-b25c-74d8d1ff5042 | easy | benjaminW2025 | 2026-08-17 20:34:45.229157+00:00 | succeeded | 2b28ce4ad54f43959e915b3381488ace022b40b62b524c6699d4b868bf69f0b9 | 7,653 | 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 = 256
NUM_HEADS = 4
... | {
"id": "286dfa92-fe8e-41c3-b25c-74d8d1ff5042",
"created_at": "2026-08-17 20:34:45.229157+00:00",
"db_md5": "a63d7312141190809f08e0bc88c77045",
"submitter": "benjawesome",
"github_login": "benjaminW2025",
"run_id": "85687b6a-8f97-4409-8e63-2606618e6e66",
"tier": "easy",
"dataset_id": "e5",
"status": "... | {
"score": {
"mean_loss": 2.2079030024770803,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.005
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.246168056411059,
"example_count": 600,
"ex... |
2870ed1a-b279-4a14-9c51-19c4958ad282 | easy | satyaki-up | 2026-08-08 10:39:03.480887+00:00 | succeeded | 964caad8bea1b26f437e5bb7089b614e401edee47b63f6ef8aa5b4ea8dc8389f | 12,097 | null | """Linear Recurrent Unit stack inspired by arXiv:2303.06349."""
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
LAYERS =... | {
"id": "2870ed1a-b279-4a14-9c51-19c4958ad282",
"created_at": "2026-08-08 10:39:03.480887+00:00",
"db_md5": "12a6ef26411a7a62f89305e936bbe153",
"submitter": "Satyaki",
"github_login": "satyaki-up",
"run_id": "0c214d22-814e-4a84-8fd5-43e7866c29a4",
"tier": "easy",
"dataset_id": "e2",
"status": "succeed... | {
"score": {
"mean_loss": 2.3414862155914307,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.002708333428017795
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.338190793991089,
"example_count": 300... |
28712037-37bb-4214-9fd4-7de44577fd20 | easy | arthurfeeney | 2026-08-07 07:08:24.458029+00:00 | succeeded | 2db1bda6fad2fd31e1cd908cd93bcfe90b5fcf400448566751dee634e05e8a3b | 12,433 | null | r"""
Notes
1. goal is basically learn y = G(x_0, T, N).
G is always a recurrence `x_t+1 = f(x_t, t) mod N`.
Shouldn't be possible in general since f isn't determinable from a dataset...
I.e., multiple f can generate the same training dataset.
2. can't really use any info on structure of f...
"""
from __future... | {
"id": "28712037-37bb-4214-9fd4-7de44577fd20",
"created_at": "2026-08-07 07:08:24.458029+00:00",
"db_md5": "0e6d32609e1d0bc89490c01ad827df7f",
"submitter": "Arthur",
"github_login": "arthurfeeney",
"run_id": "ca80ee81-5b24-47e7-abb4-b0a90d48a73a",
"tier": "easy",
"dataset_id": "e1",
"status": "succee... | {
"score": {
"mean_loss": 2.6392927169799805,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.044999999925494194
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.993319034576416,
"example_count": 100... |
28756104-7db2-45d6-9704-c5c8185a64f4 | easy | KaustubhKumar05 | 2026-08-26 01:44:27.283906+00:00 | succeeded | f23762818ce92118d01f922174102c8fa3bbc56acca27ddcd1d13de32e544e8b | 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": "28756104-7db2-45d6-9704-c5c8185a64f4",
"created_at": "2026-08-26 01:44:27.283906+00:00",
"db_md5": "0e2b265a03002721c5b7452b1fccfb5a",
"submitter": "koz",
"github_login": "KaustubhKumar05",
"run_id": "68133d6e-71b3-433b-b607-68f307ca0057",
"tier": "easy",
"dataset_id": "e6",
"status": "succee... | {
"score": {
"mean_loss": 3.7645153999328613,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.5315315574407578
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 6.036341667175293,
"example_count": 60,
... |
287765b0-dc77-4d5b-80de-a3020e1123eb | easy | yashkant | 2026-08-09 16:01:37.193955+00:00 | succeeded | 071ecc7dce420f5b8ee2d87623fa71c37f4793104f152c707b27a8448211da30 | 17,708 | null | """Round 683 fixed-depth3 AdamW LR4.5e-3 control."""
from __future__ import annotations
import math
import torch
import torch.nn.functional as F
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
TokenLossBatch,
assert_model_state,
)
from torch import Tensor, nn
... | {
"id": "287765b0-dc77-4d5b-80de-a3020e1123eb",
"created_at": "2026-08-09 16:01:37.193955+00:00",
"db_md5": "7ab29870aba37ce4e05b538d7d775d62",
"submitter": "Yash Kant",
"github_login": "yashkant",
"run_id": "b4a39be1-5296-42d8-92d5-ad3c563efe30",
"tier": "easy",
"dataset_id": "e1",
"status": "succeed... | {
"score": {
"mean_loss": 5.335506916046143,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.1483333371579647
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 7.229214191436768,
"example_count": 100,
... |
287bf37f-9e9e-47ce-8e54-76f18f761ff3 | easy | nikolageorgiev2000 | 2026-08-04 15:17:54.244763+00:00 | succeeded | 58d1d9176532ad60de5373ce659de4c7d50350d01b0bb0299c30d110817ce6f1 | 10,915 | null | """T-composed categorical-digit reasoner with fully learned pair/reduction maps."""
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,
TokenLossBat... | {
"id": "287bf37f-9e9e-47ce-8e54-76f18f761ff3",
"created_at": "2026-08-04 15:17:54.244763+00:00",
"db_md5": "a151624833dbed6e15f3f0aa771a5bc6",
"submitter": "Nikola Georgiev",
"github_login": "nikolageorgiev2000",
"run_id": "ea7836b1-796d-477f-bd13-988ba7f4d0a2",
"tier": "easy",
"dataset_id": "e2",
"s... | {
"score": {
"mean_loss": 1.255424290895462,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.7120833396911621
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 0.7539660334587097,
"example_count": 300,
... |
28855136-0541-46b1-8b83-80b922016fd0 | easy | poissonali137 | 2026-08-12 07:42:18.984312+00:00 | failed | 29a8e2f00b17aad37bf75ee795ced6a313463649efb64dbebd395c7d711e5e57 | 17,658 | null | """JordanAttention looped transformer + Muon — fully bounded loop, no norms.
Variant of jordan_loop: the loop block contains NO normalization. Stability is
structural: QK cone projection bounds scores in (0,1], sum-to-1 attention makes
the attention output a convex combination of values, and the residual state is
soft... | {
"id": "28855136-0541-46b1-8b83-80b922016fd0",
"created_at": "2026-08-12 07:42:18.984312+00:00",
"db_md5": "960af16d8b938a823b87c2881312c0fa",
"submitter": "Ali Abdul Rahim",
"github_login": "poissonali137",
"run_id": "f13c94d5-b0a8-42cb-9f94-45a2724db41e",
"tier": "easy",
"dataset_id": "e1",
"status... | null
|
28875dd7-e5f5-4f86-9c7b-a0c50745b1ec | easy | liam-gb | 2026-08-14 02:22:09.081048+00:00 | succeeded | 1d746e3ae8d56e5df34ccf93f23d12fd3fcbc2567f189b9b9955d782eab23df6 | 15,758 | null | """rns_pm_gather_ct_widet_b46rnore: 45-prime bank + per-prime temperature
+ tables at 3e-2 + reuse OFF -- every lever that has won tonight, combined.
Base: 45-prime bank + per-prime temperature + tables at lr 3e-2.
Base: widet + per-prime temperature + 45-prime bank.
The wide readout is what makes long answers reach... | {
"id": "28875dd7-e5f5-4f86-9c7b-a0c50745b1ec",
"created_at": "2026-08-14 02:22:09.081048+00:00",
"db_md5": "b8e9c96ce32883e9498e263123a1efcd",
"submitter": "liam-gb",
"github_login": "liam-gb",
"run_id": "0421b713-f66a-4082-9c35-9606b24043b8",
"tier": "easy",
"dataset_id": "e2",
"status": "succeeded"... | {
"score": {
"mean_loss": 4.921209096908569,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.4018750237300992
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 8.193352699279785,
"example_count": 300,
... |
2889beaa-2e1b-409c-b750-e6fa1b92adbf | easy | 0Chris5R | 2026-08-01 06:51:47.399526+00:00 | succeeded | 84fee3cb6b3e9bc3c429f8cbae7d60f169874ec1ef01e00bf9b3eff7561e97c4 | 22,393 | null | """Shared-digit Neural GPU with late full-objective EMA evaluation.
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 sam... | {
"id": "2889beaa-2e1b-409c-b750-e6fa1b92adbf",
"created_at": "2026-08-01 06:51:47.399526+00:00",
"db_md5": "6bd14ab4c18e4352576b1333b70aedfb",
"submitter": "Chris ",
"github_login": "0Chris5R",
"run_id": "330993e0-b292-46dd-b5ad-cea318e3849a",
"tier": "easy",
"dataset_id": "e5",
"status": "succeeded"... | {
"score": {
"mean_loss": 2.7509717416286588,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0325
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.6240947738356653,
"example_count": 600,
"... |
288a5ac3-2a19-40c0-add4-cffcf2af7adf | easy | lpbb | 2026-08-07 11:50:09.806418+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": "288a5ac3-2a19-40c0-add4-cffcf2af7adf",
"created_at": "2026-08-07 11:50:09.806418+00:00",
"db_md5": "752b8c4b53b37c5fa3e276f4404e3d02",
"submitter": "Lpbb",
"github_login": "lpbb",
"run_id": "78209975-b423-4745-b26f-19676f51b45d",
"tier": "easy",
"dataset_id": "e3",
"status": "succeeded",
"s... | {
"score": {
"mean_loss": 4.538632527890119,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.004375000009313226
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 6.050968546099055,
"example_count": 800,... |
2890e558-1191-495b-a082-b1ccaa0ed5dd | easy | erdavis0 | 2026-08-25 02:21:37.183496+00:00 | succeeded | f926cce1aea0a9d136760e9e88480e0e6166b89da0cffdae4dde2b93e05d4730 | 26,413 | 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": "2890e558-1191-495b-a082-b1ccaa0ed5dd",
"created_at": "2026-08-25 02:21:37.183496+00:00",
"db_md5": "95075821638c48bd5050ddf42eabf8c2",
"submitter": "Ethan",
"github_login": "erdavis0",
"run_id": "5ba5c3ed-07fe-4ab3-a4c8-65ea64ee11e6",
"tier": "easy",
"dataset_id": "e5",
"status": "succeeded",... | {
"score": {
"mean_loss": 4.712852876958031,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.009166666691501936
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 4.680824895610724,
"example_count": 600,... |
28923ee2-106a-4877-96d8-ccdb55697411 | easy | SunstarKK | 2026-08-18 12:06:38.993698+00:00 | succeeded | 01daa46b04c1822160f5b0bc670b694f417cfa06da13ef9fe4b504b6886a255f | 14,236 | null | """Global recurrent scratchpad for One Layer Deeper."""
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,
... | {
"id": "28923ee2-106a-4877-96d8-ccdb55697411",
"created_at": "2026-08-18 12:06:38.993698+00:00",
"db_md5": "506950446d4fbc6fe214db4273b07f6b",
"submitter": "SunstarK",
"github_login": "SunstarKK",
"run_id": "715fdbeb-e29b-408c-9ad6-d9c024d4d4f9",
"tier": "easy",
"dataset_id": "e5",
"status": "succeed... | {
"score": {
"mean_loss": 3.7852628647994244,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.00625
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 4.31229809902174,
"example_count": 600,
"e... |
28946133-4684-40a4-944e-b3e9d1bbcd76 | easy | oupadhyay | 2026-08-10 22:19:34.439011+00:00 | succeeded | 42a0d0696e57e78cde71773a4bcb65100a9dd4cf67af035b10ae113059383f88 | 6,838 | null | """Joint-register MLP candidate for the Easy tier."""
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, N_MARK, X_MARK, T_MARK, DIGIT_BASE = 0, 2, 3, 4, 7
WI... | {
"id": "28946133-4684-40a4-944e-b3e9d1bbcd76",
"created_at": "2026-08-10 22:19:34.439011+00:00",
"db_md5": "5bdbdc03af6cd568cccb6403cad02562",
"submitter": "Ojasw Upadhyay",
"github_login": "oupadhyay",
"run_id": "a6043dda-a57d-4c6a-8bab-288890523973",
"tier": "easy",
"dataset_id": "e5",
"status": "s... | {
"score": {
"mean_loss": 4.172343422167929,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.01
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 4.020514784372441,
"example_count": 600,
"exac... |
28946fa0-ba82-4f1a-9384-42309c8743da | easy | erdavis0 | 2026-08-22 11:49:08.677919+00:00 | succeeded | 489dda64c8ff330175bd25f897aab56f96331ab4e8010edd0ffa7d00d2da65bb | 9,227 | null | """Bidirectional learned Horner registers with coupled first- and second-order state."""
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,
... | {
"id": "28946fa0-ba82-4f1a-9384-42309c8743da",
"created_at": "2026-08-22 11:49:08.677919+00:00",
"db_md5": "80e0b97bc03413b4cce19f20f78b5a4b",
"submitter": "Ethan",
"github_login": "erdavis0",
"run_id": "042f8d65-75f7-4f28-999b-72744e208887",
"tier": "easy",
"dataset_id": "e10",
"status": "succeeded"... | {
"score": {
"mean_loss": 5.867223024368286,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.06905050808563828
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 4.357642650604248,
"example_count": 125,
... |
289736ba-d369-4b6f-be18-c3384c65c0c3 | easy | Yalyenea | 2026-08-18 19:05:35.056190+00:00 | succeeded | bac4dbd79938bab21f8d34f5f0e67e3385da3ce68a47dc18d378c821f9db914d | 10,300 | null | """Fixed-depth shared-transition baseline for function composition."""
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,
TokenLossBatc... | {
"id": "289736ba-d369-4b6f-be18-c3384c65c0c3",
"created_at": "2026-08-18 19:05:35.056190+00:00",
"db_md5": "11b742db94002766b35906f658560520",
"submitter": "yfff",
"github_login": "Yalyenea",
"run_id": "7b359251-7058-42bd-acba-bf817a30217a",
"tier": "easy",
"dataset_id": "e8",
"status": "succeeded",
... | {
"score": {
"mean_loss": 7.116367816925049,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0836452804505825
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 6.8814826011657715,
"example_count": 85,
... |
28a25f6a-a4a9-4c8c-b4f6-23510bfc8a77 | easy | sapient-sapiens | 2026-08-06 20:59:00.624454+00:00 | succeeded | 876449c317cc670c4ab5f6e63d747fe643c9242debc6cdfeb2577e3ad74ebe48 | 10,014 | null | """Learned Fourier arithmetic model, specialized for one-step generalization.
The network receives generic decimal values and polynomial features, then
learns its own random Fourier basis. It does not implement modular reduction
or contain a modulus-specific lookup table.
"""
from __future__ import annotations
impor... | {
"id": "28a25f6a-a4a9-4c8c-b4f6-23510bfc8a77",
"created_at": "2026-08-06 20:59:00.624454+00:00",
"db_md5": "e458ab055543d85c3f48707507ba148b",
"submitter": "Ertondy",
"github_login": "sapient-sapiens",
"run_id": "ee630674-e8ab-4d2a-bb63-e461d67c9c6a",
"tier": "easy",
"dataset_id": "e1",
"status": "su... | {
"score": {
"mean_loss": 21.54006290435791,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.20499999728053808
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 26.91628074645996,
"example_count": 100,
... |
28a399a7-8d05-473d-8fda-f62baee6a7c9 | easy | DDanlov | 2026-08-20 13:28:55.495007+00:00 | succeeded | c9e1f46496c1bf050e21b2ffd8453a4d59ac1cd5a7110f119d46d13148f17581 | 20,960 | null | from __future__ import annotations
import math
import time
from typing import Any, Optional, Tuple
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
except Imp... | {
"id": "28a399a7-8d05-473d-8fda-f62baee6a7c9",
"created_at": "2026-08-20 13:28:55.495007+00:00",
"db_md5": "656a3dfe0808abb5a670987328901f91",
"submitter": "DDanlov",
"github_login": "DDanlov",
"run_id": "6f8118d2-68e6-4bb1-8900-5dfba5aa2f46",
"tier": "easy",
"dataset_id": "e5",
"status": "succeeded"... | {
"score": {
"mean_loss": 2.671323251281687,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0141666666790843
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.672467902102278,
"example_count": 600,
... |
28a94b04-3ac7-4d7d-8383-b35b16eb5a2e | easy | DDanlov | 2026-08-14 06:57:24.407403+00:00 | succeeded | 3c8159a897485a9c230bde8949e99c5b643d6ae2e112cdba89e30250ce6a9c4b | 21,978 | 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": "28a94b04-3ac7-4d7d-8383-b35b16eb5a2e",
"created_at": "2026-08-14 06:57:24.407403+00:00",
"db_md5": "c05ed48e4644a871af1610dbf1f95d14",
"submitter": "DDanlov",
"github_login": "DDanlov",
"run_id": "b99a40d3-93fa-409b-8265-cf5fb41f3f6d",
"tier": "easy",
"dataset_id": "e5",
"status": "succeeded"... | {
"score": {
"mean_loss": 2.7172245965652846,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0141666666790843
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.719183157244071,
"example_count": 600,
... |
28aaa4e8-3e4d-4c40-9c0b-1386dd375b86 | easy | viridale | 2026-08-25 23:09:39.100451+00:00 | succeeded | b36c583d58bc25f25d3de8e03e26c4e31cff430c85cbd2985b47de7a4276e3c9 | 28,513 | null | """Constraint-posterior bilinear Transformer program.
hypothesis: v913 is three-seed T64/T64 on the single-position N law, but on
the frozen two-position law its training primal is exactly correct from step 475 while
eval is chance because unconditional position argmax deletes a required mixture. Keep
entr... | {
"id": "28aaa4e8-3e4d-4c40-9c0b-1386dd375b86",
"created_at": "2026-08-25 23:09:39.100451+00:00",
"db_md5": "b17a290751f84f487fd1dbcb406e33d9",
"submitter": "priormancer",
"github_login": "viridale",
"run_id": "bcb3f600-f2cd-4c39-81df-c51fdc88a35d",
"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... |
28b059da-e47f-40e6-9c30-b291c5168764 | easy | Alan-Qin | 2026-08-18 22:36:47.910432+00:00 | succeeded | 6e520cb7320598166811339ba27fa1056b4ce34c779a01ae5cebade1a70563d9 | 17,468 | null | """Binary-hierarchy CRT loop (v60) — solves the deep-T cold start.
Instead of one shared transition applied T times (whose gradient must survive
T compounded argmax-STEs), the model owns a small hierarchy of transition
matrices {W1, W2, W4} implementing the 1-, 2-, and 4-step maps per band. Any T
decomposes as [4]*(T/... | {
"id": "28b059da-e47f-40e6-9c30-b291c5168764",
"created_at": "2026-08-18 22:36:47.910432+00:00",
"db_md5": "40a48a800e87c38a12d4d9259034b5d4",
"submitter": "Zeyu Qin",
"github_login": "Alan-Qin",
"run_id": "0d652a17-db4a-47d0-8ad1-9a4411bb1c02",
"tier": "easy",
"dataset_id": "e7",
"status": "succeede... | {
"score": {
"mean_loss": 1.898574024438858,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.417780764400959
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.9605493545532227,
"example_count": 85,
... |
28b9268f-a5f9-4fc1-9e56-3fd81fdddad2 | easy | lpbb | 2026-08-19 12:16:39.122815+00:00 | succeeded | 7b64d46ccf85bf87ac1b5fb189826349da710a74b39f15aede2c3d4e54e475e4 | 9,817 | null | """S0 + RoPE + abacus embeddings + dropout, with a looped block: instead of
`num_layers` independent layers, a block of `num_layers` unique layers is run
`num_repeats` times with shared weights (e.g. 12 layers x 4 repeats = depth 48
but only 12 layers' worth of parameters). Weight tying across repeats forces
the block ... | {
"id": "28b9268f-a5f9-4fc1-9e56-3fd81fdddad2",
"created_at": "2026-08-19 12:16:39.122815+00:00",
"db_md5": "175cbf31ec2c524cf88af795b23a6ca6",
"submitter": "Lpbb",
"github_login": "lpbb",
"run_id": "1bf6298e-7b76-4bbd-b439-0789b501b30f",
"tier": "easy",
"dataset_id": "e1",
"status": "succeeded",
"s... | {
"score": {
"mean_loss": 5.4899609088897705,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.06833333149552345
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 6.024073123931885,
"example_count": 100,... |
28b96ea1-2b14-45ea-99c7-bb75c32b408e | easy | yashkant | 2026-08-31 07:24:54.394061+00:00 | succeeded | 67c5bd8008ef26e56d77c38c5b15710049646beca49fb41676f0b73264a065cb | 28,392 | null | """v16rrDc1 = v16rrD WITH ITS RULE-14 MACHINERY REMOVED (track a: break the fewest rules possible).
rrD's own measured record on the HIDDEN set is worse than the ancestor it was built from:
v16rr (Aug 27, submission 7f3741c9) max T 32, OOD-N max T 2
v16rrD (Aug 28, submission 11d93862) max T 8, OOD-N max T... | {
"id": "28b96ea1-2b14-45ea-99c7-bb75c32b408e",
"created_at": "2026-08-31 07:24:54.394061+00:00",
"db_md5": "6a03d453257aa73a9b4800cf8fe1c960",
"submitter": "Yash Kant",
"github_login": "yashkant",
"run_id": "c85e64c8-9cb6-4a56-9456-600fef776a14",
"tier": "easy",
"dataset_id": "e1",
"status": "succeed... | {
"score": {
"mean_loss": 15.401515007019043,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.1616666754707694
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 17.985074996948242,
"example_count": 100,... |
28c236bd-5299-44b3-8d09-78eb97490418 | easy | rteehas | 2026-08-19 11:52:09.419788+00:00 | succeeded | 21de666c2c731170cf2c7bb412d090108e7a70e6b6db191ea801398cba4c0b44 | 56,138 | null | """Anonymous marginal-program submission.
Mechanism
---------
The model is a differentiable probabilistic program. Its forward pass
computes, exactly, the answer-token predictions of a small structured latent
hypothesis space:
(prompt segmentation = which token types head the three fields)
x (field-role permutat... | {
"id": "28c236bd-5299-44b3-8d09-78eb97490418",
"created_at": "2026-08-19 11:52:09.419788+00:00",
"db_md5": "fccc28f672d2dd52219cae0c7ecd5608",
"submitter": "Ryan Teehan",
"github_login": "rteehas",
"run_id": "d7adc313-3c45-4e75-9204-15f7242fa162",
"tier": "easy",
"dataset_id": "e4",
"status": "succee... | {
"score": {
"mean_loss": 6.410785830048752,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.001875
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 6.42859121953663,
"example_count": 1200,
"... |
28c6ce5f-dacd-4a93-8f2d-c4b30b13c636 | easy | himalalps | 2026-08-07 16:57:16.681414+00:00 | succeeded | 2eeb3a3759677713be5ddb182bb5bb887a64cf88caa62eb55df08104b48d735a | 69,661 | 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": "28c6ce5f-dacd-4a93-8f2d-c4b30b13c636",
"created_at": "2026-08-07 16:57:16.681414+00:00",
"db_md5": "abfda449dc30625daf6fb0b42e5c63e4",
"submitter": "Haoyu Tang",
"github_login": "himalalps",
"run_id": "984bb8bc-2f3d-47b6-bb3f-5943253eb2cb",
"tier": "easy",
"dataset_id": "e5",
"status": "succe... | {
"score": {
"mean_loss": 3.1255368747756824,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.010833333246409893
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.3098549842834473,
"example_count": 60... |
28c72769-01d1-445c-ab5c-84420a3befd2 | easy | sapient-sapiens | 2026-08-19 18:30:46.709090+00:00 | succeeded | 8bb486c2c3b4de7de27aadf634e53f20cfd738ebbbca8f4a504918df1ed28acc | 32,686 | null | """Depth-quantized recurrence with a mixture-over-depth objective.
The recurrent state is a bank of right-aligned digit slots. Every operator
application is followed by a soft quantization back onto the token simplex, and
the digit logits produced by that quantization are the answer logits at that
depth. There is no s... | {
"id": "28c72769-01d1-445c-ab5c-84420a3befd2",
"created_at": "2026-08-19 18:30:46.709090+00:00",
"db_md5": "9ddf834c6dc18be6eb9fbe1250796b44",
"submitter": "Ertondy",
"github_login": "sapient-sapiens",
"run_id": "899d0335-fb34-4402-bb12-6d34b295726f",
"tier": "easy",
"dataset_id": "e5",
"status": "su... | {
"score": {
"mean_loss": 2.2560386328744606,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0033333333333333335
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.271840262573396,
"example_count": 60... |
28c85f14-2517-486d-90f3-c4cf7b175b63 | easy | yashkant | 2026-08-16 01:09:06.299913+00:00 | succeeded | 1a255be6e4cb64bd0db3624cd8531fb49ed367f56147b3e4a8502fdc800dc039 | 32,342 | null | """Round1779 direction-preserving radial Dykstra projection bound.
The proven Round1766 W8/D2 carrier is retained. Every learned-center
projection computes both the parent's coordinatewise tanh map and a radial
soft projection with the same asymptotic W8 vector radius. The candidate
preserves the centered direction ... | {
"id": "28c85f14-2517-486d-90f3-c4cf7b175b63",
"created_at": "2026-08-16 01:09:06.299913+00:00",
"db_md5": "29a693a613312627073e8fc1f0db8b71",
"submitter": "Yash Kant",
"github_login": "yashkant",
"run_id": "5c5704fe-9e22-4720-8c44-4fe8a8fa94da",
"tier": "easy",
"dataset_id": "e5",
"status": "succeed... | {
"score": {
"mean_loss": 2.1457962273625224,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.009583333345750967
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.1498306834644265,
"example_count": 60... |
28cb1456-302c-416c-a0a8-238867b82c1e | easy | karanganesan | 2026-08-07 08:40:58.799257+00:00 | succeeded | 78d822b7d1f9bb5b46f977e12ceb542f196ddd50f2f7a90f51b4f9012f406971 | 25,547 | 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": "28cb1456-302c-416c-a0a8-238867b82c1e",
"created_at": "2026-08-07 08:40:58.799257+00:00",
"db_md5": "c6174b599773d987b1c6e74c1c2f0956",
"submitter": "Karan Ganesan",
"github_login": "karanganesan",
"run_id": "027cdabb-1b34-4fb3-9f69-6f4deec5d953",
"tier": "easy",
"dataset_id": "e4",
"status": ... | {
"score": {
"mean_loss": 2.752418041229248,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.008020833367481828
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 3.322232961654663,
"example_count": 1200... |
28ce9f7a-02a5-450b-984a-797e59be370f | easy | jamaliki | 2026-08-07 17:10:17.705317+00:00 | succeeded | 51dc5556306a83662e6dc98a512498326d4da2fca0960a627f5f6a88bef3d40f | 98,390 | null | """One Layer Deeper submission: clocked latent loop with FSQ restoration.
Design: an encoder reads the input once into a memory. A weight-tied core is
applied m times per macro-step over [register | scratch] slots, cross-attending
to the memory (input re-injection). At each clock edge, registers are snapped
through fi... | {
"id": "28ce9f7a-02a5-450b-984a-797e59be370f",
"created_at": "2026-08-07 17:10:17.705317+00:00",
"db_md5": "5e39363b225d4cd4f906f3a56c645724",
"submitter": "Kiarash Jamali",
"github_login": "jamaliki",
"run_id": "0b85ec16-53d4-4a16-a7ef-5731919a9646",
"tier": "easy",
"dataset_id": "e1",
"status": "su... | {
"score": {
"mean_loss": 3.6759138107299805,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.08999999985098839
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 3.689039945602417,
"example_count": 100,... |
28d13e1a-5df0-4c9e-ad73-b63dd0599b67 | easy | yashkant | 2026-08-14 16:30:07.415243+00:00 | succeeded | 771b734bb8fd2c79d379ea6ae00d2d077ac22a4b0a37228b567dd012bb652192 | 31,752 | null | """Round1701 Dykstra width16/hidden32 architecture pair on the D416 MemGLU carrier.
The fixed Round1141 carrier first produces its ordinary pooled representation.
A compact primal state and two Dykstra correction memories are initialized
from that immutable representation. Every parsed public-T cycle applies two
cond... | {
"id": "28d13e1a-5df0-4c9e-ad73-b63dd0599b67",
"created_at": "2026-08-14 16:30:07.415243+00:00",
"db_md5": "87593519786b2260094f56e1f5ac1e3d",
"submitter": "Yash Kant",
"github_login": "yashkant",
"run_id": "a3553651-05a0-48c4-90f8-c9feab6c8181",
"tier": "easy",
"dataset_id": "e2",
"status": "succeed... | {
"score": {
"mean_loss": 4.724075973033905,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.2931250100955367
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 7.699827671051025,
"example_count": 300,
... |
28d50e6b-9403-4511-bfbd-e75c0bb82ea2 | easy | newjordan | 2026-08-07 00:28:13.321931+00:00 | succeeded | f1264db85e8994a858c705d18cda1aab9b0ade6a9a178fb20b22f40a7c2f5b7a | 46,081 | 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": "28d50e6b-9403-4511-bfbd-e75c0bb82ea2",
"created_at": "2026-08-07 00:28:13.321931+00:00",
"db_md5": "6cf0be9145bfc93c3b9f8747c879360d",
"submitter": "Frosty40",
"github_login": "newjordan",
"run_id": "1371e24a-bee1-4e4e-85ea-d10c61808355",
"tier": "easy",
"dataset_id": "e1",
"status": "succeed... | {
"score": {
"mean_loss": 6.856474876403809,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.04333333298563957
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 7.410281658172607,
"example_count": 100,
... |
28d74bc7-9863-4d86-8b23-2a1609679cf6 | easy | nikhilkumar26 | 2026-08-29 16:34:24.740253+00:00 | succeeded | 77f5f1fa3b4aa48dbb50f6235685cc967277e646c7d93257b334b582c3874f90 | 5,671 | null | """
One Layer Deeper Easy submission:
per-token recurrent block conditioned on a pooled N field.
Keep a hidden state at every token. Reuse one attention+MLP block.
Broadcast the N field into the MLP each loop so serial mixing can
use the modulus while digits stay separate.
No task-specific modular arithmetic solver i... | {
"id": "28d74bc7-9863-4d86-8b23-2a1609679cf6",
"created_at": "2026-08-29 16:34:24.740253+00:00",
"db_md5": "fd6b61ede7bac251c99d3e55fdb31d53",
"submitter": "Nikhil Kumar",
"github_login": "nikhilkumar26",
"run_id": "2b060169-5302-4641-84e0-0516d9029264",
"tier": "easy",
"dataset_id": "e1",
"status": ... | {
"score": {
"mean_loss": 2.458126664161682,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.014999999664723873
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.5897390842437744,
"example_count": 100... |
28d8d2d2-f416-458a-8257-4a7f02b0e2d1 | easy | DDanlov | 2026-08-26 18:31:54.793853+00:00 | failed | 2a5cc742d3e16d8fbc9f6f081c08e14c2d4c2792ed55d62c3f315878277d3ef8 | 29,496 | null | import math
import time
from typing import Any
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 ImportError:
Submission, OptimizerBundle, ModelSpec, OptimizerSpec... | {
"id": "28d8d2d2-f416-458a-8257-4a7f02b0e2d1",
"created_at": "2026-08-26 18:31:54.793853+00:00",
"db_md5": "e134c3368bbdd6780239a0ab087e6f61",
"submitter": "DDanlov",
"github_login": "DDanlov",
"run_id": "0d50f1e9-df5d-4792-8f90-8069abddc676",
"tier": "easy",
"dataset_id": "e5",
"status": "failed",
... | null
|
28da0166-65d2-4978-a45b-4e482edecf82 | easy | DDanlov | 2026-08-20 11:41:03.704046+00:00 | failed | d87708cd252ab631f9e7a9ffa02f42dd19d9fe92c377ce57463b6505b200588b | 20,764 | null | from __future__ import annotations
import math
import time
from typing import Any, Optional, Tuple
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
except Imp... | {
"id": "28da0166-65d2-4978-a45b-4e482edecf82",
"created_at": "2026-08-20 11:41:03.704046+00:00",
"db_md5": "a653119a0f5342bae5a8280387440427",
"submitter": "DDanlov",
"github_login": "DDanlov",
"run_id": "0c6cfef8-f568-487e-91f1-cfd8139170e6",
"tier": "easy",
"dataset_id": "e5",
"status": "failed",
... | null
|
28e27372-8b05-4d2d-ba30-962a1717d38e | easy | DDanlov | 2026-08-20 12:05:23.297881+00:00 | succeeded | 1aaea3eb8686a359b190c04717abf262eb42bb2cff58fe77fca0f05339887918 | 20,912 | null | from __future__ import annotations
import math
import time
from typing import Any, Optional, Tuple
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
except Imp... | {
"id": "28e27372-8b05-4d2d-ba30-962a1717d38e",
"created_at": "2026-08-20 12:05:23.297881+00:00",
"db_md5": "42c2349243501857b17d4494b8daf1fd",
"submitter": "DDanlov",
"github_login": "DDanlov",
"run_id": "3d10c338-c33e-4b88-b802-258845697898",
"tier": "easy",
"dataset_id": "e5",
"status": "succeeded"... | {
"score": {
"mean_loss": 2.3453492352864114,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.012083333345750968
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.3408979129363603,
"example_count": 60... |
28e9bd72-ae63-4acb-a332-034a483ede8d | easy | adwhit | 2026-08-15 06:08:38.716149+00:00 | succeeded | 05d0133981207a8f32d870dd1c95fb8ce5d9cfb52bf384e86351e0854bfb91fd | 3,534 | null | """Statistics-floor CONTROL, not a contender.
A model that is structurally incapable of arithmetic: its logits at position i
depend ONLY on i's offset from the RESULT marker — a learned frequency table of
"which token tends to appear at answer position k". It cannot see x, N, or T.
Whatever eval loss this reaches is t... | {
"id": "28e9bd72-ae63-4acb-a332-034a483ede8d",
"created_at": "2026-08-15 06:08:38.716149+00:00",
"db_md5": "27f0d199c676c925bfad53ce3d2e81c1",
"submitter": "adwhit",
"github_login": "adwhit",
"run_id": "f039dea8-1b26-4678-96b5-74b048ab906e",
"tier": "easy",
"dataset_id": "e2",
"status": "succeeded",
... | {
"score": {
"mean_loss": 2.214935302734375,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.014375000260770321
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.212790012359619,
"example_count": 300,... |
28ee4da2-1105-4777-bef0-4bfef8d55927 | easy | liam-gb | 2026-08-13 00:48:53.923185+00:00 | succeeded | eb2504bb7d0157e632cf799a1f18f53674416b6d2c2037173c708bf726abbc5b | 14,358 | null | """rns_pm_gather_ct_widet: dataset-sized digit head, dead loop tail trimmed.
The wide readout is what makes long answers reachable; the trim is what pays
for it, by not running loop iterations whose update is multiplied by zero.
Base docstring:
Single change vs rns_pm_gather_ct: the readout carries one
distance-from... | {
"id": "28ee4da2-1105-4777-bef0-4bfef8d55927",
"created_at": "2026-08-13 00:48:53.923185+00:00",
"db_md5": "1a548bea555b987cf0a175795dc3aa83",
"submitter": "liam-gb",
"github_login": "liam-gb",
"run_id": "0801002a-a45d-4264-9dd6-d30b5325b33f",
"tier": "easy",
"dataset_id": "e3",
"status": "succeeded"... | {
"score": {
"mean_loss": 9.72793436050415,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.10749999899417162
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 10.989352226257324,
"example_count": 800,
... |
28f22ee1-8e5d-451c-8d60-953f403c381f | easy | mnida | 2026-08-23 21:42:19.947399+00:00 | succeeded | 485181a5dc2cac5cae35d96dbdcaae501e655ed6998dc225ab909666754c4ffe | 10,001 | null | """Canonical recurrent workspace with a learned discrete depth selector."""
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,
assert_model_... | {
"id": "28f22ee1-8e5d-451c-8d60-953f403c381f",
"created_at": "2026-08-23 21:42:19.947399+00:00",
"db_md5": "eb140ab9cb196569dc3db6f53ea9b016",
"submitter": "mnida",
"github_login": "mnida",
"run_id": "3a0c681c-ff57-4cbc-8579-78f5dd752f22",
"tier": "easy",
"dataset_id": "e6",
"status": "succeeded",
... | {
"score": {
"mean_loss": 2.0523234009742737,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.04009009199216962
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 1.9829906225204468,
"example_count": 60,... |
28f8f399-f59f-417e-8b27-24e28c47913f | easy | yashkant | 2026-08-16 00:10:26.221479+00:00 | succeeded | c0648b7098c1872e6b3375ce22f742e00036bcce9f0888550d29a8d01e387c57 | 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": "28f8f399-f59f-417e-8b27-24e28c47913f",
"created_at": "2026-08-16 00:10:26.221479+00:00",
"db_md5": "8d5bfd0a9d59e2753f5da607f3d8895f",
"submitter": "Yash Kant",
"github_login": "yashkant",
"run_id": "e35dfc3f-16e2-4747-a333-04760e35ed1f",
"tier": "easy",
"dataset_id": "e2",
"status": "succeed... | {
"score": {
"mean_loss": 3.0833535194396973,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.07312500476837158
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 3.9896111488342285,
"example_count": 300... |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.