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 |
|---|---|---|---|---|---|---|---|---|---|---|
0ac6aeae-dc07-4420-ac62-eacf2513db04 | easy | shirvani-jr | 2026-08-29 02:53:40.074763+00:00 | succeeded | 459da493dd2908cad9a8c9872fbad420462944f11d66ff0b4106740ae7ffd3e9 | 47,739 | null | """Adaptive learned discrete recurrent fabric for One Layer Deeper.
A gradient-trained, operation-free recurrent computational machine. The
model learns, from the competition's endpoint supervision: how to parse the
prompt into a bank of categorical digit registers, a reusable transition
circuit applied serially, a l... | {
"id": "0ac6aeae-dc07-4420-ac62-eacf2513db04",
"created_at": "2026-08-29 02:53:40.074763+00:00",
"db_md5": "bd4bfcc15bf65f683c4a1d57d23e89cb",
"submitter": "Ali",
"github_login": "shirvani-jr",
"run_id": "43331cab-39e1-454a-b43f-457b1a1548fa",
"tier": "easy",
"dataset_id": "e5",
"status": "succeeded"... | {
"score": {
"mean_loss": 2.3793831362582005,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.005416666711680591
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.3820762634277344,
"example_count": 60... |
0aca2268-bd7e-4254-9eee-1ea1401075fe | easy | sapient-sapiens | 2026-08-15 21:21:52.188934+00:00 | succeeded | 469ae48173121f8a328b5c1f44d0bff0e2f098d243dcd227975cee12399b6829 | 17,772 | null | """T-only adaptive chunk routing over a learned shared operator.
The controller sees only the categorical tokens inside the T prompt field and
is trained solely by final-answer loss. It never parses T numerically and T is
not used as an intermediate supervision target. Inputs use learned additive
token, semantic-gro... | {
"id": "0aca2268-bd7e-4254-9eee-1ea1401075fe",
"created_at": "2026-08-15 21:21:52.188934+00:00",
"db_md5": "38cda1d694d318221e86f82dd23bbc39",
"submitter": "Ertondy",
"github_login": "sapient-sapiens",
"run_id": "610636f5-4de7-460a-ab73-13971a16046d",
"tier": "easy",
"dataset_id": "e2",
"status": "su... | {
"score": {
"mean_loss": 2.287271457919482,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.004791666666666667
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.3041355111457746,
"example_count": 300... |
0ad80aaf-7e41-423b-9c14-086d82d36b3d | easy | adamsoliev | 2026-08-01 23:08:09.395932+00:00 | failed | 4e5e218ebc7924a0b4573afddcc33a08452821c3b304702453f04e37d0f71054 | 35,723 | null | """Dual-factor semigroup-block automaton for One Layer Deeper."""
from __future__ import annotations
import math
import torch
import torch.nn.functional as F
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
assert_model_state,
)
from torch import Tensor, nn
WIDTH = ... | {
"id": "0ad80aaf-7e41-423b-9c14-086d82d36b3d",
"created_at": "2026-08-01 23:08:09.395932+00:00",
"db_md5": "6522259187ade0b39126e1487cbbcf74",
"submitter": "Adam Soliev",
"github_login": "adamsoliev",
"run_id": "d7ec4388-15c3-409d-8a11-8707124b10d3",
"tier": "easy",
"dataset_id": "e4",
"status": "fai... | null
|
0ad88eed-1c7e-4b35-8c84-3d76973408c1 | easy | viridale | 2026-08-15 03:58:50.001679+00:00 | succeeded | 60e4873f59d8ec5482d9a5547cc74b76f0d2827fd1faafd72cbdaaededf2b36b | 7,887 | null | """Clean autonomous recurrence with small random tied initialization.
hypothesis: v167's default unit-variance embedding creates CE 335 through the tied
decoder, while untying overfits and fixed scaling cannot sharpen; standard 0.02
random token/position initialization should retain tying with ordinary initial... | {
"id": "0ad88eed-1c7e-4b35-8c84-3d76973408c1",
"created_at": "2026-08-15 03:58:50.001679+00:00",
"db_md5": "5078e3dc2de717e587541b93a51960b3",
"submitter": "priormancer",
"github_login": "viridale",
"run_id": "de939991-4fa3-413d-9215-08d43b2aacf9",
"tier": "easy",
"dataset_id": "e6",
"status": "succe... | {
"score": {
"mean_loss": 4.271920442581177,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.08040541037917137
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 4.03049373626709,
"example_count": 60,
... |
0add292f-acc0-464a-8b5a-a2a993c697e2 | easy | gauravmishra | 2026-08-08 05:58:36.344523+00:00 | succeeded | 75b766be3bdb202352ea93a6e16ba69e78f53670d40d6dfd8e80bb467030a160 | 3,787 | null | """Four-block Transformer stack reused for two complete cycles."""
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_... | {
"id": "0add292f-acc0-464a-8b5a-a2a993c697e2",
"created_at": "2026-08-08 05:58:36.344523+00:00",
"db_md5": "e5cb4f65a2ccf87b2b6b891df6235afe",
"submitter": "Gaurav Mishra",
"github_login": "gauravmishra",
"run_id": "331e42f2-718b-4967-97f0-aeb801667c2f",
"tier": "easy",
"dataset_id": "e5",
"status": ... | {
"score": {
"mean_loss": 10.11096706084038,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.005416666666666667
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 10.523661686165985,
"example_count": 600... |
0ae0c155-5109-4c11-88e5-392cbf8e5367 | easy | sapient-sapiens | 2026-08-17 19:07:48.908901+00:00 | succeeded | 6cea0e1561f811c6befb625391b41eaa02da14c043493e59e0a70c06fde3281b | 28,659 | null | """v40 lineage: T-free operator, N memory, PonderNet mix, no T in the loss.
Field-local encoder, T-free core, countable halt, answer-slot readout.
N is seeded and restored every step. T is never parsed in forward or loss.
"""
from __future__ import annotations
import math
import time
import torch
import torch.nn.fu... | {
"id": "0ae0c155-5109-4c11-88e5-392cbf8e5367",
"created_at": "2026-08-17 19:07:48.908901+00:00",
"db_md5": "5809d68726f93c793f4740bae499a5d9",
"submitter": "Ertondy",
"github_login": "sapient-sapiens",
"run_id": "f0876079-0759-4f8e-92c0-72704fb5494e",
"tier": "easy",
"dataset_id": "e9",
"status": "su... | {
"score": {
"mean_loss": 1.8729384304294059,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.013333333432674408
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 1.8712642192840576,
"example_count": 90... |
0aeabc3f-3cfb-453a-9360-a68fdd877da5 | easy | sapient-sapiens | 2026-08-31 01:58:06.206578+00:00 | succeeded | 550e0ecfaa5cbfabe794c8d00fa409304423802acfa6f94be2070fb3ebf1f66b | 17,896 | null | """Eligible E5 live canonical-X operator with categorical boundary matching.
This is the strict S-to-A, live canonical-X parent with only its learned
endpoint selector replaced by exact categorical matching in the existing
task/digit chart. Every Horner digit is read from the current square operand;
after each complet... | {
"id": "0aeabc3f-3cfb-453a-9360-a68fdd877da5",
"created_at": "2026-08-31 01:58:06.206578+00:00",
"db_md5": "2bf9c29f128bc3a0e5318b05ca8a3f4a",
"submitter": "Ertondy",
"github_login": "sapient-sapiens",
"run_id": "105e7c4d-3095-49b7-bd1b-7a3eefa7d3c5",
"tier": "easy",
"dataset_id": "e5",
"status": "su... | {
"score": {
"mean_loss": 2.1391535147666976,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.005
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.1387393752555677,
"example_count": 600,
"e... |
0af47151-4d50-4f33-b773-498ba2ae3bb1 | easy | karanganesan | 2026-08-04 01:36:12.864751+00:00 | succeeded | a2a6a52febad15170cc09f132592c0b881e3662c6d1acc0808786b9f15e5e47f | 11,993 | 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": "0af47151-4d50-4f33-b773-498ba2ae3bb1",
"created_at": "2026-08-04 01:36:12.864751+00:00",
"db_md5": "59a0513e316bcd4bf0f0a6b66c911b28",
"submitter": "Karan Ganesan",
"github_login": "karanganesan",
"run_id": "07e203d6-ab8c-4656-8c17-a08057b95b0d",
"tier": "easy",
"dataset_id": "e1",
"status": ... | {
"score": {
"mean_loss": 2.405043601989746,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.03833333356305957
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.147099256515503,
"example_count": 100,
... |
0b02a667-56a7-43a5-b4ef-2765a74250f1 | easy | oupadhyay | 2026-08-11 08:53:12.513879+00:00 | succeeded | ff8cf80a01c629621e4e770e9d5b4bf11c551c0a861b344e6b546c27ccf3608e | 8,092 | null | """Attention-augmented 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... | {
"id": "0b02a667-56a7-43a5-b4ef-2765a74250f1",
"created_at": "2026-08-11 08:53:12.513879+00:00",
"db_md5": "47833f79fd8b2e1ee32e658cb10acac2",
"submitter": "Ojasw Upadhyay",
"github_login": "oupadhyay",
"run_id": "0965de09-c48a-4d65-ae1e-8f4d246d93bb",
"tier": "easy",
"dataset_id": "e5",
"status": "s... | {
"score": {
"mean_loss": 2.5242650687665353,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.00375
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.4344778103678753,
"example_count": 600,
... |
0b2691ac-354c-4e99-bb8c-af098896874b | easy | erdavis0 | 2026-08-26 04:46:24.120050+00:00 | succeeded | 34958bd1fecf39e0cd177cec37f4a5b763f56ecda181924fc21973bf094acd2c | 30,362 | null | """Latent-role continuous REV C3: sequence mixture plus role balance.
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 and f... | {
"id": "0b2691ac-354c-4e99-bb8c-af098896874b",
"created_at": "2026-08-26 04:46:24.120050+00:00",
"db_md5": "63a88d8095c75c1574375cc20d89bc96",
"submitter": "Ethan",
"github_login": "erdavis0",
"run_id": "ab11b82e-d78f-4770-9781-14957df79fe1",
"tier": "easy",
"dataset_id": "e10",
"status": "succeeded"... | {
"score": {
"mean_loss": 6.361447215080261,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.18862627632915974
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 3.8654563426971436,
"example_count": 125,... |
0b295ce0-c1b0-4fba-87c3-fa72b66806ff | easy | khushidahi | 2026-08-03 02:58:16.149646+00:00 | succeeded | 1d11730430b1849bf4300611bd82a4b2b374b937422b2f61fb5c30e610e77d12 | 14,072 | 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": "0b295ce0-c1b0-4fba-87c3-fa72b66806ff",
"created_at": "2026-08-03 02:58:16.149646+00:00",
"db_md5": "ef1d9e1eaee18948547e9dd3a651cb0c",
"submitter": "khushidahi",
"github_login": "khushidahi",
"run_id": "c6c24861-7030-4870-9f05-0168e1b4c902",
"tier": "easy",
"dataset_id": "e5",
"status": "succ... | {
"score": {
"mean_loss": 2.2029902935028076,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0054166666232049465
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.201563596725464,
"example_count": 60... |
0b2a6860-1111-4a84-a9f6-b6e839073a52 | easy | ehonig | 2026-08-14 06:00:56.806405+00:00 | succeeded | 0b5619ef3c7504cb18e5285debd2e18fe694c43cf834dded2e12e38c62418560 | 52,555 | null | """DigitLoop — a place-value recurrent model for repeated modular squaring.
The task is ``y = x^(2^T) mod N`` with the answer read back tail aligned, one
decimal place value per token position from the right. A stack of attention and
MLP layers has no primitive for the two operations this needs — a partial-product
co... | {
"id": "0b2a6860-1111-4a84-a9f6-b6e839073a52",
"created_at": "2026-08-14 06:00:56.806405+00:00",
"db_md5": "c15463c498305d470b6e2d5be8015b79",
"submitter": "Edouardo Honig",
"github_login": "ehonig",
"run_id": "15fb8b9b-2486-454d-90f0-b6b6c3c9b0f8",
"tier": "easy",
"dataset_id": "e1",
"status": "succ... | {
"score": {
"mean_loss": 4.331276535987854,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.4750000089406967
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 3.2555787563323975,
"example_count": 100,
... |
0b31935a-1956-4c1f-bd05-11db02128792 | easy | DDanlov | 2026-08-17 07:56:23.321552+00:00 | succeeded | 4cd7d6ad801af8bb45d196a24e62efc955557e863ae33fb36a25f6a6f37b9d03 | 27,486 | 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": "0b31935a-1956-4c1f-bd05-11db02128792",
"created_at": "2026-08-17 07:56:23.321552+00:00",
"db_md5": "1e8dba91b7e16a4c2b7b702c64886fac",
"submitter": "DDanlov",
"github_login": "DDanlov",
"run_id": "c9287ea9-9ac2-4a34-a599-750544e31f49",
"tier": "easy",
"dataset_id": "e1",
"status": "succeeded"... | {
"score": {
"mean_loss": 3.494648076649616,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.029999999652306236
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 4.437749862670898,
"example_count": 100,... |
0b370948-fb5d-4c53-8d66-fc7411cec0e6 | easy | msnow0 | 2026-08-08 11:27:36.235861+00:00 | succeeded | a2cc31a109a27abdc5ae774e4996e7c518610b87d55fb63aff0a287f5ff9fbb0 | 17,246 | null | """One Layer Deeper submission — mechanism 1 (see BRIEF.md §6, research/RULES.md).
Weight-tied recurrent operator over a discrete token tape. One outer iteration per
squaring step (loop count conditioned on the prompt's T, which is explicitly allowed);
between outer steps the latent is snapped back onto token embeddin... | {
"id": "0b370948-fb5d-4c53-8d66-fc7411cec0e6",
"created_at": "2026-08-08 11:27:36.235861+00:00",
"db_md5": "dedd4e1fa347c6ff57b720bf470975ea",
"submitter": "msnow0",
"github_login": "msnow0",
"run_id": "00ffa6f3-9e26-49a0-ba39-fc3b62c913cb",
"tier": "easy",
"dataset_id": "e1",
"status": "succeeded",
... | {
"score": {
"mean_loss": 4.442922353744507,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.036666667088866234
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 4.406261920928955,
"example_count": 100,... |
0b373309-f48f-4293-ae07-a8c427f3f7c6 | easy | adambrzosko | 2026-08-30 17:31:26.327017+00:00 | succeeded | 8781610f7ae4a775c53e9903cfafb95b69d4d847b3b5ef5af1ec26229732259d | 12,489 | null | """One Layer Deeper — digit-slot recurrent machine, v2.
Prompt: [N] digits(N) [X] digits(x) [T] digits(T); label: the decimal expansion
of x^(2^T) mod N, tail-aligned, so prompt position len-1-s carries the digit of
significance s.
The model parses the three fields onto a grid of slots indexed by (role,
significance)... | {
"id": "0b373309-f48f-4293-ae07-a8c427f3f7c6",
"created_at": "2026-08-30 17:31:26.327017+00:00",
"db_md5": "0dea9ab432f016cae8b8c9d38ffb9d23",
"submitter": "Adam Brzosko",
"github_login": "adambrzosko",
"run_id": "ec96cf21-ce09-40e4-b9ee-9c7b2a39ae1a",
"tier": "easy",
"dataset_id": "e5",
"status": "s... | {
"score": {
"mean_loss": 3.044873999348729,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.010833333333333334
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.4386591013236965,
"example_count": 600... |
0b4063f0-a9cb-4590-8788-02ccba214e87 | easy | pculaf | 2026-08-25 14:23:57.276180+00:00 | succeeded | 597baf57582cf77e3dc7f2c91edc50c3075aadb145afd41399e0e44b7759c3c8 | 8,287 | null | """Four-loop post-RMSNorm Transformer for One Layer Deeper."""
from __future__ import annotations
from collections.abc import Iterable
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
assert_m... | {
"id": "0b4063f0-a9cb-4590-8788-02ccba214e87",
"created_at": "2026-08-25 14:23:57.276180+00:00",
"db_md5": "b018b35b15360a396a53c7a098f345e9",
"submitter": "Pavle Culafic",
"github_login": "pculaf",
"run_id": "31578ae4-3e68-4ce9-bf98-07a9ce639519",
"tier": "easy",
"dataset_id": "e10",
"status": "succ... | {
"score": {
"mean_loss": 2.052752733230591,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0025252525229007006
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.052879571914673,
"example_count": 125... |
0b413cdc-538a-4413-8700-1a22157c2965 | easy | adamsoliev | 2026-08-01 22:31:32.696605+00:00 | succeeded | 138a2ea4df36969b58362e66bbcb17b1b2f75856511532425bef8c42f5b28fcb | 31,457 | null | """Paired-step factor-coordinate automaton for One Layer Deeper."""
from __future__ import annotations
import math
import torch
import torch.nn.functional as F
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
assert_model_state,
)
from torch import Tensor, nn
WIDTH ... | {
"id": "0b413cdc-538a-4413-8700-1a22157c2965",
"created_at": "2026-08-01 22:31:32.696605+00:00",
"db_md5": "6827f036de8a4989ea87e5e2a0f96c6e",
"submitter": "Adam Soliev",
"github_login": "adamsoliev",
"run_id": "bac11e7a-66e8-4526-815a-8700ea4c810e",
"tier": "easy",
"dataset_id": "e5",
"status": "suc... | {
"score": {
"mean_loss": 0.5321380296025434,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.9475000051657358
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 0.6442761782038907,
"example_count": 600,... |
0b440bbb-1383-4d2a-9800-80c6c9263fe5 | easy | rteehas | 2026-08-19 12:01:08.989110+00:00 | succeeded | 28045ee699e7ebb27d64b0c08ce12e91ea593c3ee7b2093707e19fbbab7e1a19 | 54,458 | 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": "0b440bbb-1383-4d2a-9800-80c6c9263fe5",
"created_at": "2026-08-19 12:01:08.989110+00:00",
"db_md5": "926cd5e9ceec4d487d17a5b68185bbc7",
"submitter": "Ryan Teehan",
"github_login": "rteehas",
"run_id": "0509cb7f-692d-4101-919f-3d372754aca8",
"tier": "easy",
"dataset_id": "e2",
"status": "succee... | {
"score": {
"mean_loss": 2.833213448524475,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.8332133293151855,
"example_count": 300,
"exac... |
0b5eb171-2b1f-4a2a-b7fc-be6b69a072bb | easy | briheet | 2026-08-05 11:28:14.611351+00:00 | succeeded | 01e90237ff14f594a6c9bee9d1b09560cd820441c441541ed328f4fd4ed7e5ba | 4,156 | 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 = 96
NUM_HEADS = 4
NU... | {
"id": "0b5eb171-2b1f-4a2a-b7fc-be6b69a072bb",
"created_at": "2026-08-05 11:28:14.611351+00:00",
"db_md5": "1247b5363fc9fbf771970449381c1ef3",
"submitter": "Aizenn",
"github_login": "briheet",
"run_id": "8c77d98e-6f59-44b5-9e39-8204c7634ab7",
"tier": "easy",
"dataset_id": "e1",
"status": "succeeded",... | {
"score": {
"mean_loss": 2.873664975166321,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.04999999888241291
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 3.1112325191497803,
"example_count": 100,... |
0b6643d4-e959-4aea-bcd5-0f7b86e7c656 | easy | sapient-sapiens | 2026-08-09 15:22:53.780857+00:00 | succeeded | 82f21506979a8a5d9c0eb433f0ea37acb56c0718b095ed4a32fc2ee278c14862 | 14,774 | null | """Near-500M adaptive recurrence grown from the hosted-working baseline."""
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,
... | {
"id": "0b6643d4-e959-4aea-bcd5-0f7b86e7c656",
"created_at": "2026-08-09 15:22:53.780857+00:00",
"db_md5": "84e8f375bc70efa26f3758b837f26381",
"submitter": "Ertondy",
"github_login": "sapient-sapiens",
"run_id": "42130def-8be4-4c72-a88c-5c54f0d877bd",
"tier": "easy",
"dataset_id": "e1",
"status": "su... | {
"score": {
"mean_loss": 2.423716397560357,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.2765538848158138,
"example_count": 100,
"exac... |
0b6f0a2b-03aa-4e9f-a29d-73fb3ec56615 | easy | MarcoF1 | 2026-08-13 16:45:14.781245+00:00 | succeeded | fbe213b38fad5332ced983962119ff0670c520f28a36e11de082b55e17c1ea24 | 14,397 | null | """Residue-clock model for fixed-N generalization.
Representation: parsed integer values of x and N feed (a) exact residue-phase
features over a bank of small integer periods, (b) learned double-hashed value
embeddings (per-residue identity), and T feeds cyclic phase features
(T mod 2..48) plus saturating channels — p... | {
"id": "0b6f0a2b-03aa-4e9f-a29d-73fb3ec56615",
"created_at": "2026-08-13 16:45:14.781245+00:00",
"db_md5": "498f495d8a0b910efb7ca633ca25eee2",
"submitter": "Marco Fleming",
"github_login": "MarcoF1",
"run_id": "f99c6f13-386f-4993-90c7-d1083c6aad6a",
"tier": "easy",
"dataset_id": "e1",
"status": "succ... | {
"score": {
"mean_loss": 2.9394758250053385,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.08500000034769375
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 3.4677839279174805,
"example_count": 100... |
0b732ace-a3ba-4ff5-a7ff-0fdc01916039 | easy | chad-atexpedient | 2026-08-27 04:08:34.471376+00:00 | succeeded | 79edd83086734adbfe6635b8c5f5efd0ecffb930572fdfd01bbe746ee0866ba5 | 20,128 | null | """X139: carry-aware multiplicative coupling.
X138 proved a same-place value*value / value*modulus coupling lifts bilateral
Medium T1 (seen 3, OOD-N 1). A true product also couples adjacent places through
carry. X139 extends the coupling input with adjacent-place products
(value_center * value_lower and value_center *... | {
"id": "0b732ace-a3ba-4ff5-a7ff-0fdc01916039",
"created_at": "2026-08-27 04:08:34.471376+00:00",
"db_md5": "5c9bcae73a3f0f918970c6ff13a45496",
"submitter": "chad-atexpedient",
"github_login": "chad-atexpedient",
"run_id": "36ce3f6e-78de-4c4c-98b4-6a166d71f2cd",
"tier": "easy",
"dataset_id": "e5",
"st... | {
"score": {
"mean_loss": 2.246638546088615,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.010833333345750967
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.2882582232556534,
"example_count": 600... |
0b73402d-c07e-4d07-917d-a4458433db9e | easy | lpbb | 2026-08-21 13:25:04.850725+00:00 | succeeded | ad2978d4986c6fc71c0b244ee1247c866f346a077f7907166e215f94f7e6218c | 12,545 | null | """S0 + RoPE + abacus + looped block, with *input-adaptive* depth: instead of
a fixed `num_repeats`, each sample's T is parsed from its own token stream
(the digits after the `T` marker) and the shared block of `num_layers` layers
is applied exactly T times for that sample. One block pass = one modular
squaring step, s... | {
"id": "0b73402d-c07e-4d07-917d-a4458433db9e",
"created_at": "2026-08-21 13:25:04.850725+00:00",
"db_md5": "c59d29e73b5a6e9505ae111d7ef2b646",
"submitter": "Lpbb",
"github_login": "lpbb",
"run_id": "7d83e0ec-68d9-4044-8cd5-39cdc08cfa88",
"tier": "easy",
"dataset_id": "e1",
"status": "succeeded",
"s... | {
"score": {
"mean_loss": 5.958294868469238,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.041666666977107525
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 5.763108730316162,
"example_count": 100,... |
0b74e09d-af49-49b3-94e0-052249c6f9fc | easy | oupadhyay | 2026-08-08 11:17:08.190808+00:00 | failed | c3fdf581ea85e1584cddf9a3e102a458a8707a7b38f96ff11f4f7c21848a521b | 9,569 | null | """Tied equilibrium transition over categorical, LSD-relative digit places."""
from __future__ import annotations
import math
import torch
from torch import Tensor, nn
import torch.nn.functional as F
from benchmark import ModelSpec, OptimizerBundle, OptimizerSpec, Submission, assert_model_state
PAD, N_MARK, X_MARK, ... | {
"id": "0b74e09d-af49-49b3-94e0-052249c6f9fc",
"created_at": "2026-08-08 11:17:08.190808+00:00",
"db_md5": "c66f4472db28e8d6ba96413a9f309f5a",
"submitter": "Ojasw Upadhyay",
"github_login": "oupadhyay",
"run_id": "eb98f202-d3ee-4a69-93f6-798acf3007e6",
"tier": "easy",
"dataset_id": "e5",
"status": "f... | null
|
0b782cde-3321-46f0-94d3-f475ecdaff07 | easy | blake-camp-surge | 2026-08-03 03:18:36.233998+00:00 | succeeded | 6d8926dfe7898e55ff56cda6a63a7262340b6a72e930feb58d307e2817360899 | 10,100 | null | """Field-aware, weight-tied recurrent model for modular composition."""
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_stat... | {
"id": "0b782cde-3321-46f0-94d3-f475ecdaff07",
"created_at": "2026-08-03 03:18:36.233998+00:00",
"db_md5": "046f405bfc06960528f520347448215e",
"submitter": "Blake Camp",
"github_login": "blake-camp-surge",
"run_id": "fd173e11-0602-4da2-a0e3-e85d11e47346",
"tier": "easy",
"dataset_id": "e5",
"status":... | {
"score": {
"mean_loss": 6.347001527100579,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0070833333457509685
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 7.064180720547389,
"example_count": 600... |
0b7a1f3e-c26f-471c-89e4-1504b82dcfec | easy | erdavis0 | 2026-08-19 08:10:33.652058+00:00 | succeeded | 7ef166017dce10b8421c57ae518af1658d74b82dfaf1497b4512d7154140a749 | 9,290 | null | """Raw-vocabulary recurrent operator with identity state decoding."""
from __future__ import annotations
import math
import time
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
TokenLossBatch... | {
"id": "0b7a1f3e-c26f-471c-89e4-1504b82dcfec",
"created_at": "2026-08-19 08:10:33.652058+00:00",
"db_md5": "e44becb8207f8dd37af235cdc45557ce",
"submitter": "Ethan",
"github_login": "erdavis0",
"run_id": "0a6c7eb1-7230-41bc-8afb-f1e11c08a344",
"tier": "easy",
"dataset_id": "e3",
"status": "succeeded",... | {
"score": {
"mean_loss": 2.270950426688847,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.018125000000000002
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.270412677769707,
"example_count": 800,... |
0b7acfa3-8d5f-47a2-960a-2f2a6ca0da86 | easy | jordanrubin | 2026-08-09 02:39:26.684312+00:00 | succeeded | 23c96b25d2a5b82c71ed8ca6cf4fadcc3fc10ccfe43a16949682eade745a366c | 16,637 | 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": "0b7acfa3-8d5f-47a2-960a-2f2a6ca0da86",
"created_at": "2026-08-09 02:39:26.684312+00:00",
"db_md5": "550ddc9f9d2598d880f0a082c3722567",
"submitter": "Jordan Rubin",
"github_login": "jordanrubin",
"run_id": "5d39b448-afc6-4d9e-bbd3-4ebcd7f8ac82",
"tier": "easy",
"dataset_id": "e5",
"status": "s... | {
"score": {
"mean_loss": 3.3707728777210955,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.005416666545594731
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.2205818759218996,
"example_count": 60... |
0b7cd0fc-5228-4425-b46b-c7c4a2e0156c | easy | shirvani-jr | 2026-08-29 00:37:18.888533+00:00 | succeeded | 155f606ce031d624b8d7c65ad1d99f90e5bb944b4018f3110d2c404510e6b42e | 49,680 | null | """Adaptive learned discrete recurrent fabric for One Layer Deeper.
A gradient-trained, operation-free recurrent computational machine. The
model learns, from the competition's endpoint supervision: how to parse the
prompt into a bank of categorical digit registers, a reusable transition
circuit applied serially, a l... | {
"id": "0b7cd0fc-5228-4425-b46b-c7c4a2e0156c",
"created_at": "2026-08-29 00:37:18.888533+00:00",
"db_md5": "cb3afeb71ed15e08a5efb721bfd7329f",
"submitter": "Ali",
"github_login": "shirvani-jr",
"run_id": "f1060ee7-9d01-422b-83d4-4421d9bda3c2",
"tier": "easy",
"dataset_id": "e1",
"status": "succeeded"... | {
"score": {
"mean_loss": 2.421492278575897,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.3699999898672104
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 3.458063840866089,
"example_count": 100,
... |
0b7deff8-4cf0-4876-af51-5d2015998245 | easy | 0Chris5R | 2026-08-01 06:54:50.004688+00:00 | succeeded | 481ea59ee8f0df07976a4574e9c41fbd54e7767fd4e8022a5dcd3d2a3d37fe71 | 20,589 | null | """Shared-digit Neural GPU with a compute-matched operator curriculum.
The model aligns decimal digits by significance, forms a learned low-rank
second-order feature map of the current state, and evolves a two-dimensional
workspace with the original convolutional-GRU equations. Each arithmetic
transition applies the ... | {
"id": "0b7deff8-4cf0-4876-af51-5d2015998245",
"created_at": "2026-08-01 06:54:50.004688+00:00",
"db_md5": "bf3c781c63adc7c025f6fe47ba83e811",
"submitter": "Chris ",
"github_login": "0Chris5R",
"run_id": "9a7f86f9-d9bc-4e54-a2a7-52bf2478326c",
"tier": "easy",
"dataset_id": "e5",
"status": "succeeded"... | {
"score": {
"mean_loss": 3.364802680965619,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.018750000012417632
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 3.392775689539888,
"example_count": 600,... |
0b80556f-65de-46b2-b9c9-98c60d1de700 | easy | neel04 | 2026-08-11 13:05:54.918812+00:00 | failed | ea4f6391b3cd377c97a17de3493d9236734def90592e181cfc805b11a5837492 | 17,362 | null | """Layer 6: stationary NAG with tied decoding and an auxiliary critic."""
from __future__ import annotations
import math
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
BatchReuseContext,
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
... | {
"id": "0b80556f-65de-46b2-b9c9-98c60d1de700",
"created_at": "2026-08-11 13:05:54.918812+00:00",
"db_md5": "dd890b6d242e892d1c71f2b2860aa360",
"submitter": "Neel Gupta",
"github_login": "neel04",
"run_id": "a5958afe-3239-466d-aeb4-45ddf545c7f0",
"tier": "easy",
"dataset_id": "e1",
"status": "failed",... | null
|
0b82a70b-3879-4d77-9f21-4c76fd5cdfff | easy | isaac0804 | 2026-08-04 21:54:41.508887+00:00 | succeeded | be7806836d1cea71d7f3e2f07ec58a794d6b18ac097bcf44ec8bc5ce7dc3e99e | 8,903 | null | """Weight-tied looped Transformer combining the two things learned from
looped_transformer_undamped and looped_transformer_curriculum separately:
- No damped gate: h_{t+1} = block(h_t) directly (looped_transformer_undamped
showed damping's ~12%-per-step update is why the loop's T=1 rows never
consolidated the sing... | {
"id": "0b82a70b-3879-4d77-9f21-4c76fd5cdfff",
"created_at": "2026-08-04 21:54:41.508887+00:00",
"db_md5": "799c89b7c07fe2048176f7feb17e0d3d",
"submitter": "Isaac Yong",
"github_login": "isaac0804",
"run_id": "ee274cab-4c43-42a0-8b02-1ade33f5e128",
"tier": "easy",
"dataset_id": "e1",
"status": "succe... | {
"score": {
"mean_loss": 2.182335615158081,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.024999999441206455
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 1.9974517822265625,
"example_count": 100... |
0b8e3d1f-6918-487a-af44-667bf6da6f60 | easy | ehonig | 2026-08-19 17:18:57.218463+00:00 | succeeded | fd39a8caa9796ffe701380ad16c9d439ce469ec577bf106ae5168e26a914aa26 | 18,780 | 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": "0b8e3d1f-6918-487a-af44-667bf6da6f60",
"created_at": "2026-08-19 17:18:57.218463+00:00",
"db_md5": "cff6a14fca5bc6e4f02e293d4e1f4221",
"submitter": "Edouardo Honig",
"github_login": "ehonig",
"run_id": "1e6e8598-c8d0-4ea6-a87b-e4c890536110",
"tier": "easy",
"dataset_id": "e3",
"status": "succ... | {
"score": {
"mean_loss": 2.2352139949798584,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.016874999273568392
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.2479333877563477,
"example_count": 80... |
0b945da8-b582-43ad-924b-981efb23e21c | easy | k-penchev | 2026-08-22 23:18:11.134003+00:00 | succeeded | 450bc14d47b28e1ec66f7b80a98dd1f7e15aa2176c755fb320f857024ae6b750 | 8,441 | null | """Gated width-96 recurrence optimized with self-contained Schedule-Free AdamW."""
from __future__ import annotations
from collections.abc import Iterable
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Subm... | {
"id": "0b945da8-b582-43ad-924b-981efb23e21c",
"created_at": "2026-08-22 23:18:11.134003+00:00",
"db_md5": "448720dadd0d64a27da871562f6b887b",
"submitter": "Kaloyan Penchev",
"github_login": "k-penchev",
"run_id": "05b08d3d-0bfb-461c-a9a8-3f09c34b8d23",
"tier": "easy",
"dataset_id": "e1",
"status": "... | {
"score": {
"mean_loss": 2.2706247568130493,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.07999999821186066
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.5516297817230225,
"example_count": 100... |
0b96099e-fa41-48d0-bd90-8e22830d0651 | easy | ehonig | 2026-08-14 01:13:19.401915+00:00 | succeeded | bed7b46c989217f150e8f32594973831db97cd435e186415766922012faf890c | 48,079 | null | """DigitLoop — a place-value recurrent model for repeated modular squaring.
The task is ``y = x^(2^T) mod N`` with the answer read back tail aligned, one
decimal place value per token position from the right. A stack of attention and
MLP layers has no primitive for the two operations this needs — a partial-product
co... | {
"id": "0b96099e-fa41-48d0-bd90-8e22830d0651",
"created_at": "2026-08-14 01:13:19.401915+00:00",
"db_md5": "8a71b19028930edd1dd4facf7e650670",
"submitter": "Edouardo Honig",
"github_login": "ehonig",
"run_id": "46097bef-eef5-425e-9d44-8a8318e47acd",
"tier": "easy",
"dataset_id": "e1",
"status": "succ... | {
"score": {
"mean_loss": 6.065410733222961,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.20999999344348907
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 8.629045486450195,
"example_count": 100,
... |
0ba0e2ea-2028-4eb4-a757-4f855617a0e6 | easy | khushidahi | 2026-08-07 04:17:22.680951+00:00 | succeeded | 0125a353d0618fd1034432dd47902c152ac902c697ed07f2b729a5b4ea1dde80 | 16,538 | 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": "0ba0e2ea-2028-4eb4-a757-4f855617a0e6",
"created_at": "2026-08-07 04:17:22.680951+00:00",
"db_md5": "57ac5147ed11c09bc16e18b0bb5a4897",
"submitter": "khushidahi",
"github_login": "khushidahi",
"run_id": "2aa87cfc-b7a8-4c23-8b58-7f83d6cf16ac",
"tier": "easy",
"dataset_id": "e5",
"status": "succ... | {
"score": {
"mean_loss": 2.1523882150650024,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.007916666800156236
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.1509485244750977,
"example_count": 60... |
0bbadfe5-b6f6-4963-b863-23cffd31627a | easy | ehonig | 2026-08-21 01:02:09.989196+00:00 | succeeded | a9ad0d4126ab74869f8e7f4109841541ffaa4c10fe840580936c9852aab9df05 | 30,222 | 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": "0bbadfe5-b6f6-4963-b863-23cffd31627a",
"created_at": "2026-08-21 01:02:09.989196+00:00",
"db_md5": "e6d7ec5bd1fddf2d609c6fc1ca4d18f7",
"submitter": "Edouardo Honig",
"github_login": "ehonig",
"run_id": "79b6af7d-2c87-4358-9fe0-58cd6c61f4d5",
"tier": "easy",
"dataset_id": "e1",
"status": "succ... | {
"score": {
"mean_loss": 2.0220086574554443,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.07999999821186066
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.0389645099639893,
"example_count": 100... |
0bc82ef4-345f-4108-be00-0d1736dcd76a | easy | DDanlov | 2026-08-17 19:12:48.901741+00:00 | succeeded | 85637adb2dc699e21a0d7885dce88f9af735a5f2d173c8cbfb13de967bd4fb9f | 23,669 | 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": "0bc82ef4-345f-4108-be00-0d1736dcd76a",
"created_at": "2026-08-17 19:12:48.901741+00:00",
"db_md5": "85da6d3e58ecc5e0a8311e24ea101322",
"submitter": "DDanlov",
"github_login": "DDanlov",
"run_id": "6af09c64-87ef-42e9-b593-27e3a4d9fb35",
"tier": "easy",
"dataset_id": "e1",
"status": "succeeded"... | {
"score": {
"mean_loss": 16.18325285309546,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.028333333370586238
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 25.33155342358262,
"example_count": 100,... |
0bcccb0f-697c-4b44-975f-db9547dbd77b | easy | thedhruvrawat | 2026-08-19 15:43:11.104311+00:00 | succeeded | 538b2e101b8b04a166d6a153a67b0edbe37ce4a1ed04c6dd5386f06bfbbffcb3 | 21,868 | null | from __future__ import annotations
import inspect
import math
import time
from typing import Any, Sequence
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
assert_model_state,
)
WIDTH = 384
NU... | {
"id": "0bcccb0f-697c-4b44-975f-db9547dbd77b",
"created_at": "2026-08-19 15:43:11.104311+00:00",
"db_md5": "1ef95d293d5d5b5e46ec901b3beedcf3",
"submitter": "Dhruv Rawat",
"github_login": "thedhruvrawat",
"run_id": "0a76912a-c333-45f3-bf24-8187219e34ea",
"tier": "easy",
"dataset_id": "e1",
"status": "... | {
"score": {
"mean_loss": 6.533103942871094,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.03999999910593033
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 6.957754611968994,
"example_count": 100,
... |
0bd00453-a1b8-455a-adde-11b1ee33a641 | easy | khushidahi | 2026-08-08 02:45:02.898511+00:00 | succeeded | f9fc94b1787cb1698014730b178d529beb7febb7178c5d6a79338cf49e72c707 | 15,587 | null | """R42 Fourier-number representation experiment.
All input-dependent numeric parsing stays inside differentiable GPU tensor
operations. This is a learned-model representation, not a task solver: it does
not compute modular squaring or create intermediate labels.
"""
from __future__ import annotations
import math
impor... | {
"id": "0bd00453-a1b8-455a-adde-11b1ee33a641",
"created_at": "2026-08-08 02:45:02.898511+00:00",
"db_md5": "b13f80d9638d44c0e9d71db2ca47dc1e",
"submitter": "khushidahi",
"github_login": "khushidahi",
"run_id": "e2de3107-7e4c-4d48-b10e-9c33f5b8ec07",
"tier": "easy",
"dataset_id": "e5",
"status": "succ... | {
"score": {
"mean_loss": 2.1701444291115344,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.006666666707023978
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.1775572299957275,
"example_count": 60... |
0bd28245-08ca-42de-acb2-7ebb7c05cb13 | easy | DDanlov | 2026-08-15 21:06:34.926316+00:00 | succeeded | 3eb517cb9667863720ce75d439fd67ebc36b79a26534e0b145dad12284313550 | 17,950 | null | """
Official Submission for One Layer Deeper Challenge
Architecture: Deep Equilibrium Port-Hamiltonian Deliberation Network (Model C v3 / DEPH-Net)
- Input: Multi-Scale Linear Harmonic Torus Frequency Encoder (omega_j = 1..64)
- Recurrent Core: Continuous Port-Hamiltonian Dynamical Flow dz/dt = (J - R) * grad_E
* J =... | {
"id": "0bd28245-08ca-42de-acb2-7ebb7c05cb13",
"created_at": "2026-08-15 21:06:34.926316+00:00",
"db_md5": "623970f9d216ffb45b2ef4f8507de941",
"submitter": "DDanlov",
"github_login": "DDanlov",
"run_id": "16d66386-289e-492a-8485-40387c62c65e",
"tier": "easy",
"dataset_id": "e5",
"status": "succeeded"... | {
"score": {
"mean_loss": 2.4435929158256338,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0033333333581686024
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.4521222737337975,
"example_count": 6... |
0bdd527c-79ae-47e9-bea3-caa9b4285195 | easy | gauravmishra | 2026-08-10 03:30:02.419513+00:00 | succeeded | 0d513986ce04dab666731f5e907f56704fc2cc487f258e3ae665a3a014636603 | 6,788 | null | """E14: full-clock U4 with OrthoGrad-projected AdamW gradients.
OrthoGrad projection adapted from the official
LucasPrietoAl/grokking-at-the-edge-of-numerical-stability commit
720d2444df12b851d6cb417ab08cf125c822b2ae under the MIT License.
Copyright (c) 2025 Lucas Prieto
Permission is hereby granted, free of charge, t... | {
"id": "0bdd527c-79ae-47e9-bea3-caa9b4285195",
"created_at": "2026-08-10 03:30:02.419513+00:00",
"db_md5": "8b8b6354fcd64dcad312767cfd5f7191",
"submitter": "Gaurav Mishra",
"github_login": "gauravmishra",
"run_id": "16188d30-ff0c-4f74-b6bc-69267eaf044f",
"tier": "easy",
"dataset_id": "e5",
"status": ... | {
"score": {
"mean_loss": 5.166237276096982,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.00625
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 5.336726986239309,
"example_count": 600,
"e... |
0bde2126-6baf-48b6-8175-3e56ab7d9b45 | easy | erdavis0 | 2026-08-24 16:16:23.885422+00:00 | succeeded | 0f22f44dcea1ea5e06251ffc0d25d191c75b949b4fee5c94d2cc22a8908e7873 | 6,407 | null | """Low-rank prompt-conditioned operator over right-aligned slots."""
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
RANK ... | {
"id": "0bde2126-6baf-48b6-8175-3e56ab7d9b45",
"created_at": "2026-08-24 16:16:23.885422+00:00",
"db_md5": "354e0e6275589d14b11c0ef21e053498",
"submitter": "Ethan",
"github_login": "erdavis0",
"run_id": "817c74d7-c891-4bd2-855e-22841d46d2bf",
"tier": "easy",
"dataset_id": "e2",
"status": "succeeded",... | {
"score": {
"mean_loss": 11.423971176147461,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.008958333870396018
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 11.914266586303711,
"example_count": 30... |
0bdea11f-861b-45c2-bc82-af5a72a9b5e7 | easy | ehonig | 2026-08-20 10:23:26.488869+00:00 | succeeded | 4ea294ea13100087c4d92f4ce7ce6a8266bf946c4d3f31dd54bf380c2eee70bd | 24,999 | 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": "0bdea11f-861b-45c2-bc82-af5a72a9b5e7",
"created_at": "2026-08-20 10:23:26.488869+00:00",
"db_md5": "2ee5f47eff2c098a2fe5b509e2eba126",
"submitter": "Edouardo Honig",
"github_login": "ehonig",
"run_id": "2ec9cd01-6fff-40fb-816a-79549d31276b",
"tier": "easy",
"dataset_id": "e1",
"status": "succ... | {
"score": {
"mean_loss": 2.184258818626404,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0833333320915699
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.2005960941314697,
"example_count": 100,
... |
0be3bd12-7a3c-4900-9f90-25588ce67bb7 | easy | liam-gb | 2026-08-17 01:50:50.588867+00:00 | succeeded | 24fc878b19987564e88eccf1d98937ae5735fa8ab620fdef470b9eacfec14e2f | 16,740 | null | """rns_pm_gather_ct_widet_l32nore_r8: single change vs
rns_pm_gather_ct_widet_l32nore — the batch-reuse policy returns True, so the
evaluator reuses each fetched batch for the maximum eight optimizer updates.
WHY. On the granular Easy suites one batch is 85-97% of the training split,
so every step ends an epoch and th... | {
"id": "0be3bd12-7a3c-4900-9f90-25588ce67bb7",
"created_at": "2026-08-17 01:50:50.588867+00:00",
"db_md5": "46c5decfff7d97a850fb297981e9d310",
"submitter": "liam-gb",
"github_login": "liam-gb",
"run_id": "7661d74e-2240-467e-9b1a-6328997bda73",
"tier": "easy",
"dataset_id": "e1",
"status": "succeeded"... | {
"score": {
"mean_loss": 3.3711744248867035,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.5616666823625565
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 5.832922458648682,
"example_count": 100,
... |
0be4ce90-c91c-4a11-9c0c-e7734cf2634a | easy | karanganesan | 2026-08-08 04:12:45.396071+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": "0be4ce90-c91c-4a11-9c0c-e7734cf2634a",
"created_at": "2026-08-08 04:12:45.396071+00:00",
"db_md5": "3e8439ba8ab561733e7dc883fab4ecb2",
"submitter": "Karan Ganesan",
"github_login": "karanganesan",
"run_id": "5916c3a9-968c-40b5-a86c-b07d125084b2",
"tier": "easy",
"dataset_id": "e3",
"status": ... | {
"score": {
"mean_loss": 8.618232011795044,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.004999999771825969
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 9.291633605957031,
"example_count": 800,... |
0be79d3f-fe03-4fd4-a7ab-6f589359b8b2 | easy | falloficaruss | 2026-08-06 05:40:06.384434+00:00 | succeeded | ba4b5e879e5f8b3b5ce3b042f1b1bfe31cdce345706ed6b733b434bfb5fce6bf | 12,041 | null | """Place-aware deep encoder v6.
e3 evidence
-----------
looped v4 (00fd/3068): 27ms/step, loss stuck ~2.05, exact ~0.6% — no algorithm.
deep untied v5 (600bf3b7): 39ms/step, loss 2.83→1.27, train exact 19%, test 0.4%.
Learning is real but still mostly train-set fit in 60s.
v6
--
* decimal place-from-right features ... | {
"id": "0be79d3f-fe03-4fd4-a7ab-6f589359b8b2",
"created_at": "2026-08-06 05:40:06.384434+00:00",
"db_md5": "630487996c79f37909539f8db42bfe4c",
"submitter": "Abhishek Shinde",
"github_login": "falloficaruss",
"run_id": "84c0cbf2-abfc-4212-b828-2d944635fb74",
"tier": "easy",
"dataset_id": "e3",
"status... | {
"score": {
"mean_loss": 2.757644685398007,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.005000000149011611
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 3.017839406386212,
"example_count": 800,... |
0bf493f4-a068-41f7-892f-d69639bae52f | easy | nikolageorgiev2000 | 2026-08-04 14:52:11.129719+00:00 | succeeded | 2c247befd3ac77c7eeea4011ba5a6ebe0bbb262b6415f11439e9517ff557aa65 | 10,923 | 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": "0bf493f4-a068-41f7-892f-d69639bae52f",
"created_at": "2026-08-04 14:52:11.129719+00:00",
"db_md5": "523f243cc70c800ccb8e7b46d110a951",
"submitter": "Nikola Georgiev",
"github_login": "nikolageorgiev2000",
"run_id": "1172f8fd-5d3f-4eb4-b3f8-a0de3e881bc7",
"tier": "easy",
"dataset_id": "e1",
"s... | {
"score": {
"mean_loss": 1.7123360335826874,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.7050000131130219
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 0.7953526377677917,
"example_count": 100,... |
0bf4c499-30b3-41b4-9d85-b4d1644478b0 | easy | achyutreddy24 | 2026-08-03 18:44:50.935049+00:00 | succeeded | 071f514bf4eb0dc7d2c803d51e7ce7b595141fbe6a62afb396e6fe232579349d | 14,574 | null | """Fuel-gated recurrent residue clock (informed initialization) — public submission.
WHAT THIS MODEL IS
==================
The task is x^(2^T) mod N from decimal digit tokens in a single parallel forward
pass. This submission computes it the way the algebra suggests: as T composed
applications of one exact squaring s... | {
"id": "0bf4c499-30b3-41b4-9d85-b4d1644478b0",
"created_at": "2026-08-03 18:44:50.935049+00:00",
"db_md5": "ebfdda4b212b209bb459b9984a40b1d7",
"submitter": "Achyut Reddy",
"github_login": "achyutreddy24",
"run_id": "2eaef510-5076-4449-97a6-76e41bc04661",
"tier": "easy",
"dataset_id": "e1",
"status": ... | {
"score": {
"mean_loss": 0.7278791666030884,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 1.0
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 0.7278791666030884,
"example_count": 100,
"exa... |
0bf6b173-b8e6-4021-857b-64eb666296f4 | easy | sapient-sapiens | 2026-08-09 14:52:30.880999+00:00 | succeeded | 1763607be036fe83d25121757c497c2d5b7d375786d04e914da63b7224eb45c8 | 11,416 | null | """Near-500M adaptive recurrence grown from the hosted-working baseline."""
from __future__ import annotations
import math
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import ModelSpec, OptimizerBundle, OptimizerSpec, Submission, assert_model_state
D_MODEL = 6400
NUM_HE... | {
"id": "0bf6b173-b8e6-4021-857b-64eb666296f4",
"created_at": "2026-08-09 14:52:30.880999+00:00",
"db_md5": "7975d62f91a0f4a130ee0cdd08af97db",
"submitter": "Ertondy",
"github_login": "sapient-sapiens",
"run_id": "3b2b2949-b1ab-439f-9a90-cf6873d8dd29",
"tier": "easy",
"dataset_id": "e5",
"status": "su... | {
"score": {
"mean_loss": 2.4555802936374684,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.005
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.455016273888237,
"example_count": 600,
"ex... |
0bf7e380-24d1-49d9-be49-05bb042ed640 | easy | DDanlov | 2026-08-15 20:57:37.388928+00:00 | succeeded | 3eb517cb9667863720ce75d439fd67ebc36b79a26534e0b145dad12284313550 | 17,950 | null | """
Official Submission for One Layer Deeper Challenge
Architecture: Deep Equilibrium Port-Hamiltonian Deliberation Network (Model C v3 / DEPH-Net)
- Input: Multi-Scale Linear Harmonic Torus Frequency Encoder (omega_j = 1..64)
- Recurrent Core: Continuous Port-Hamiltonian Dynamical Flow dz/dt = (J - R) * grad_E
* J =... | {
"id": "0bf7e380-24d1-49d9-be49-05bb042ed640",
"created_at": "2026-08-15 20:57:37.388928+00:00",
"db_md5": "623970f9d216ffb45b2ef4f8507de941",
"submitter": "DDanlov",
"github_login": "DDanlov",
"run_id": "12cd5ad7-79cf-468c-a32c-c6cf0497fd17",
"tier": "easy",
"dataset_id": "e1",
"status": "succeeded"... | {
"score": {
"mean_loss": 2.3640678929937393,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.05166666684051355
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.240490263967372,
"example_count": 100,... |
0c0aa8e6-d42f-41f4-9612-6c988aa0776a | easy | ddoan | 2026-08-14 12:27:58.781040+00:00 | succeeded | 944d94ef7be950b53c704800b7a31baa34185c98dab8e985c0f2bc2a25e858eb | 9,939 | null | """Pinned numeral codec with a directly supervised tied transition."""
from __future__ import annotations
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
BatchReuseContext,
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
TokenLossBatch,... | {
"id": "0c0aa8e6-d42f-41f4-9612-6c988aa0776a",
"created_at": "2026-08-14 12:27:58.781040+00:00",
"db_md5": "f6b9295c8aba251144a38135ba6c1444",
"submitter": "Doug Doan",
"github_login": "ddoan",
"run_id": "ec8e6fbc-08ae-458a-92b8-0b3f4fc3004c",
"tier": "easy",
"dataset_id": "e5",
"status": "succeeded"... | {
"score": {
"mean_loss": 2.2821011310363897,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.005833333395421505
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.350017264284895,
"example_count": 600... |
0c0cddaa-0a14-464d-a6d8-72ff91eb1325 | easy | chad-atexpedient | 2026-08-20 15:33:16.796910+00:00 | succeeded | b97caeca21bd5b4261323a708154740e903d1a3aa15233fe07cb79ffba4d6d40 | 22,936 | null | """X114: pair-count-normalized ordered routed self-convolution.
X105 sums every pair on an output anti-diagonal, so route magnitude changes with
digit length and destination. Although RMSNorm removes global scale, summed noise
and gradient variance still grow with pair count. X114 divides each contraction
by the squar... | {
"id": "0c0cddaa-0a14-464d-a6d8-72ff91eb1325",
"created_at": "2026-08-20 15:33:16.796910+00:00",
"db_md5": "da58411c6517642347acb9d1d98e67a9",
"submitter": "chad-atexpedient",
"github_login": "chad-atexpedient",
"run_id": "1dcf4b6e-f5a0-4973-8f8b-4909c4693d49",
"tier": "easy",
"dataset_id": "e5",
"st... | {
"score": {
"mean_loss": 2.4260369704175657,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.008333333333333333
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.518561983322349,
"example_count": 600... |
0c0f4e47-5280-497c-a4d4-f597cf1009b4 | easy | Dandandan | 2026-08-05 17:27:13.990628+00:00 | succeeded | a8fcf5b5cdfa710a62db673ef2c810e7cf252a3a0c07ab4b20918329166b0640 | 34,142 | 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": "0c0f4e47-5280-497c-a4d4-f597cf1009b4",
"created_at": "2026-08-05 17:27:13.990628+00:00",
"db_md5": "102be411c18dd63338ae33c1b6c0cd40",
"submitter": "Daniël Heres",
"github_login": "Dandandan",
"run_id": "0697ccdb-9c6d-4389-a2bf-ccb6726b8b99",
"tier": "easy",
"dataset_id": "e5",
"status": "suc... | {
"score": {
"mean_loss": 1.1348570754328067e-05,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 1.0
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 1.1365176222235958e-05,
"example_count": 600,
... |
0c13f752-eb91-4b93-9704-c65458c7c505 | easy | sapient-sapiens | 2026-08-31 01:33:43.470868+00:00 | succeeded | b35d8b9191a06a3e94314d6ecdb3d2b22ca88208bd32a4116a82fb92d8f69374 | 17,243 | null | """Eligible E5 direct-Horner model with a shared two-block stack per write.
Every latent-S write and every A commit passes through the same ordered pair of
Transformer blocks. The stack is tied across all four Horner positions and
all eight squares, giving deeper local computation without stage-specialized
weights. ... | {
"id": "0c13f752-eb91-4b93-9704-c65458c7c505",
"created_at": "2026-08-31 01:33:43.470868+00:00",
"db_md5": "fcfef7149eb9edbbbb95f1f1ff88895c",
"submitter": "Ertondy",
"github_login": "sapient-sapiens",
"run_id": "a0876b11-c497-4a52-86df-89066428a29d",
"tier": "easy",
"dataset_id": "e5",
"status": "su... | {
"score": {
"mean_loss": 2.186361524152763,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.008750000012417634
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.190483114644551,
"example_count": 600,... |
0c1476ea-9faa-4ff6-af4d-d6d881a4186f | easy | christopherhwood | 2026-08-08 03:52:26.987085+00:00 | succeeded | c69aa0c55ea42808dc2afb2cdffe05967a3cc1e1cc4d4fcdb20df1e074bde063 | 19,055 | null | """Canonical categorical cellular program for a hidden recurrence.
The model turns the positive endpoint-identification argument into an explicit
architecture. The state that survives every local update contains only a
distribution over the ten observable decimal digits and a small categorical
controller. One transla... | {
"id": "0c1476ea-9faa-4ff6-af4d-d6d881a4186f",
"created_at": "2026-08-08 03:52:26.987085+00:00",
"db_md5": "def6c82a877fe2d828f883372b27d1b1",
"submitter": "Chris Wood",
"github_login": "christopherhwood",
"run_id": "4e7749a0-e63b-4aec-9ecc-991c7cb0c04c",
"tier": "easy",
"dataset_id": "e5",
"status":... | {
"score": {
"mean_loss": 2.2154825524375186,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.005833333333333334
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.211072805216494,
"example_count": 600... |
0c1551ba-7613-4990-9acf-f0cd3f8576ca | easy | erdavis0 | 2026-08-29 01:12:46.867285+00:00 | succeeded | c30171170d2368108bbd1f8de018dfbd1d387e0e28dec2d543ae91ea35b7ad59 | 20,136 | null | """Curvature table with a GPU-resident cached modulus-relative residual."""
from __future__ import annotations
import math
import time
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
TokenLos... | {
"id": "0c1551ba-7613-4990-9acf-f0cd3f8576ca",
"created_at": "2026-08-29 01:12:46.867285+00:00",
"db_md5": "efb4da441ead29fd5826581673cba669",
"submitter": "Ethan",
"github_login": "erdavis0",
"run_id": "85a37bd6-5a23-4143-85c6-75a23845eeab",
"tier": "easy",
"dataset_id": "e9",
"status": "succeeded",... | {
"score": {
"mean_loss": 0.03999961285444442,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.9900000095367432
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 0.00023136523668654263,
"example_count":... |
0c1cb830-f6a0-49b6-a62a-db1866ab9665 | easy | johnowhitaker | 2026-08-03 21:40:19.222076+00:00 | succeeded | aee1152277d06400ee29607fe9ac7c6dc3f0a5793108acb2c2bf10a5d09ef131 | 11,943 | null | """Interleaved residue-state recurrent Transformer.
Four possible residue positions at the right edge of each prompt receive five
internal spacing slots apiece. A shared Transformer body is rerun six times.
After each pass, the state at the current prediction position is inserted into
the next spacing slot, creating ... | {
"id": "0c1cb830-f6a0-49b6-a62a-db1866ab9665",
"created_at": "2026-08-03 21:40:19.222076+00:00",
"db_md5": "9c6a71dcba1d6ca357eaaea4f41ac08c",
"submitter": "Jonathan Whitaker",
"github_login": "johnowhitaker",
"run_id": "1230d5ca-40c2-4547-8962-65aef017991d",
"tier": "easy",
"dataset_id": "e1",
"stat... | {
"score": {
"mean_loss": 1.9397597312927246,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.07999999821186066
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 1.929322600364685,
"example_count": 100,... |
0c1dbbe8-4e28-487e-8571-e8bacd571034 | easy | DDanlov | 2026-08-14 20:06:14.517601+00:00 | succeeded | 1e2ebbf158c7e4118b50fad00f4b51fc3d174bc8e2497924e6bddef8d86c44d1 | 22,526 | 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": "0c1dbbe8-4e28-487e-8571-e8bacd571034",
"created_at": "2026-08-14 20:06:14.517601+00:00",
"db_md5": "ffedd3ab2e873ea7ee169c7550909429",
"submitter": "DDanlov",
"github_login": "DDanlov",
"run_id": "83969121-15a3-44d3-8579-843be676304f",
"tier": "easy",
"dataset_id": "e3",
"status": "succeeded"... | {
"score": {
"mean_loss": 2.37196415617022,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.018125000000000002
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.3739731535472237,
"example_count": 800,... |
0c1edb7f-a831-46d7-80ed-7fe92f55aabf | easy | oupadhyay | 2026-08-26 15:32:32.476046+00:00 | failed | 564c6a1f88f8e72e72dd0c9153fbc9fa1f6e6dc90e8bb60c0565db7bf77ea540 | 4,094 | null | """Persistent dynamic-width cell with a literal learned polynomial update."""
from __future__ import annotations
import torch
import torch.nn.functional as F
from torch import nn
from benchmark import OptimizerBundle,Submission,TokenLossBatch,assert_model_state
D,H=56,28
class C:
def __init__(s,vocab_size,max_seq_len)... | {
"id": "0c1edb7f-a831-46d7-80ed-7fe92f55aabf",
"created_at": "2026-08-26 15:32:32.476046+00:00",
"db_md5": "7e78ff3211f9672beadab6d9d3e1ac0d",
"submitter": "Ojasw Upadhyay",
"github_login": "oupadhyay",
"run_id": "27ddf923-0fbf-44e1-9a52-487b20e9f623",
"tier": "easy",
"dataset_id": "e5",
"status": "f... | null
|
0c1f137c-4cfd-454c-9153-904737fae965 | easy | oupadhyay | 2026-08-22 17:37:28.100348+00:00 | succeeded | 386786fb26ad2fda8f40738e5beff99a13f93d639c454869d0d74be821b204cc | 9,193 | null | """D64 three-sweep chunkwise DeltaNet transition for the Easy benchmark."""
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
PAD, N_MARK, X_MARK, T_M... | {
"id": "0c1f137c-4cfd-454c-9153-904737fae965",
"created_at": "2026-08-22 17:37:28.100348+00:00",
"db_md5": "197ac9c259babf6bb851293085202b82",
"submitter": "Ojasw Upadhyay",
"github_login": "oupadhyay",
"run_id": "a89cfdc3-08dd-4928-8036-59c4489c099b",
"tier": "easy",
"dataset_id": "e5",
"status": "s... | {
"score": {
"mean_loss": 2.3644382827387567,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.01041666670391957
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.2891828447179408,
"example_count": 600... |
0c2e1385-927c-4517-b67b-98519c37b912 | easy | chad-atexpedient | 2026-08-09 20:06:22.770815+00:00 | succeeded | bdd68e9076ad80269fdd7fb662a0a9ea0edb83413bc5e8b635e2db27dfb4b86f | 9,918 | null | """BL4B: BL1 strong-decay control."""
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
IMMUTABLE_LANES = 3... | {
"id": "0c2e1385-927c-4517-b67b-98519c37b912",
"created_at": "2026-08-09 20:06:22.770815+00:00",
"db_md5": "535316953d45ff2eba215b1135302f40",
"submitter": "chad-atexpedient",
"github_login": "chad-atexpedient",
"run_id": "e9e491e6-3453-471c-9306-3d01955ae3d5",
"tier": "easy",
"dataset_id": "e5",
"st... | {
"score": {
"mean_loss": 5.8001245714163305,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.004583333333333333
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 5.770626686079085,
"example_count": 600... |
0c5bade6-3c5b-4566-b6ae-c3c86fe64799 | easy | benjaminW2025 | 2026-08-17 07:24:49.121949+00:00 | succeeded | 12cb7b1530bf8d6834004a9783a4f1233424799ab4a809e0c5c0c5bdcc0296fc | 3,664 | 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": "0c5bade6-3c5b-4566-b6ae-c3c86fe64799",
"created_at": "2026-08-17 07:24:49.121949+00:00",
"db_md5": "07dd19f9c21d332bf663f1cb359860dd",
"submitter": "benjawesome",
"github_login": "benjaminW2025",
"run_id": "42c3e51f-54c7-40e8-95eb-7f28a419e67d",
"tier": "easy",
"dataset_id": "e1",
"status": "... | {
"score": {
"mean_loss": 2.8160598278045654,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.009999999776482582
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 3.480668544769287,
"example_count": 100... |
0c5f6cd3-e2af-49c6-a826-f4f22abe11d8 | easy | G-AshwinKumar | 2026-08-13 16:34:02.935319+00:00 | succeeded | 7bf30e19d90e7f6cadb4ad67d0b2cd51634a4fd2d29a507be6f840a0dd78f8c1 | 3,628 | null | """Basic single-pass Transformer with PyTorch AdamW."""
from __future__ import annotations
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
assert_model_state,
)
D_MODEL = 128
NUM_HEADS = 4
... | {
"id": "0c5f6cd3-e2af-49c6-a826-f4f22abe11d8",
"created_at": "2026-08-13 16:34:02.935319+00:00",
"db_md5": "eaac69fe345f82973fcf392c504a40ff",
"submitter": "Ashwin Kumar",
"github_login": "G-AshwinKumar",
"run_id": "e3719d27-b727-4443-8092-45516f6578ff",
"tier": "easy",
"dataset_id": "e1",
"status": ... | {
"score": {
"mean_loss": 3.089873790740967,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 4.115736484527588,
"example_count": 100,
"exact... |
0c738643-8619-4c01-a9d6-b75c9203c7d6 | easy | int3 | 2026-08-22 16:55:39.203118+00:00 | succeeded | e400baaabe292fb856f67a0930282c56e9af2b28c4531c250061d2c5d0f3c74b | 23,718 | null | """ALU v10e: v10d with 14-digit registers (40-bit N) and confirmed elites.
The model executes iterated modular arithmetic exactly, using learned
carry/borrow tables (initialized from a chosen distribution, per the
organizers' ruling) applied in the eval-hard/argmax regime as integer table
lookups. All values live as p... | {
"id": "0c738643-8619-4c01-a9d6-b75c9203c7d6",
"created_at": "2026-08-22 16:55:39.203118+00:00",
"db_md5": "13c13b69a1833f5cfd56efc73f236bd3",
"submitter": "Jez Ng",
"github_login": "int3",
"run_id": "0b26ed37-3881-4062-975e-c00f6b14f626",
"tier": "easy",
"dataset_id": "e5",
"status": "succeeded",
... | {
"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": 600,
"exact_accuracy": 1.0,
"... |
0c741765-73b3-468b-9834-43b8af9c15bd | easy | dlyr3 | 2026-08-23 20:07:33.582929+00:00 | failed | f94b635f6dd2b70c17e8a3d599ceaea226881d1a6201c012187cb986cfacb760 | 7,386 | null | """The M-layer of arXiv:2008.03936 applied to the repeated-squaring task.
M = B + T(U x) is built from the pooled prompt, exp(M) is taken with
torch.matrix_exp, and V + S exp(M) is read back out. The paper's activity regularisation is deliberately NOT applied: measured on e1,
any nonzero lambda drives ||exp(M)||_F... | {
"id": "0c741765-73b3-468b-9834-43b8af9c15bd",
"created_at": "2026-08-23 20:07:33.582929+00:00",
"db_md5": "3f569f9edd87fb3629c5fd5274b314cf",
"submitter": "dlyr3",
"github_login": "dlyr3",
"run_id": "52ec77f0-3fb7-4fbb-8b87-f106abc6e0c8",
"tier": "easy",
"dataset_id": "e1",
"status": "failed",
"st... | null
|
0c7e710e-64eb-4313-a16a-0d97502d4cdd | easy | akkolli | 2026-08-07 02:00:04.003426+00:00 | succeeded | 5da46b60e3d67b15319455f8a020cfce711240ab81ca6f55c01a2f793bf2e772 | 189,886 | null | """Ordinal-codebook square lane plus one learned affine-square recurrence.
The ten digit-token categories receive trainable scalar scores under each of
two serialization hypotheses. They start in the participant-visible canonical
token order, which makes fixed-modulus datasets well posed, but remain fully
trainable s... | {
"id": "0c7e710e-64eb-4313-a16a-0d97502d4cdd",
"created_at": "2026-08-07 02:00:04.003426+00:00",
"db_md5": "4e4f93faa40a3dd5d69e508a0a49d987",
"submitter": "Akshay Kolli",
"github_login": "akkolli",
"run_id": "3d2ad6b2-6805-4e5a-a055-3300ab142ac9",
"tier": "easy",
"dataset_id": "e3",
"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": 800,
"exact_accuracy": 1.0,
"... |
0c7f917a-5210-4612-b2c1-0198fc298d5c | easy | jordanrubin | 2026-08-02 18:17:13.433178+00:00 | succeeded | 4e035996a92f3747a56b144314883b965da03ed01a2a2edd4428ae1b9d2d89b2 | 13,994 | 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": "0c7f917a-5210-4612-b2c1-0198fc298d5c",
"created_at": "2026-08-02 18:17:13.433178+00:00",
"db_md5": "1ed8cd827d4f4241972ae9f1e82cca7e",
"submitter": "Jordan Rubin",
"github_login": "jordanrubin",
"run_id": "e8594986-131b-49e1-be60-e14556bb4d67",
"tier": "easy",
"dataset_id": "e1",
"status": "s... | {
"score": {
"mean_loss": 1.9084710335045543,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 1.8724752308717414,
"example_count": 100,
"exa... |
0c81a9a9-f848-4a56-a7e7-f20796628c95 | easy | oupadhyay | 2026-08-07 14:52:37.347134+00:00 | succeeded | 4a790af9f35063d404c069d0f33f9623a284fa41e3b027cb1f9110af83b5b1a6 | 6,742 | null | """Easy E5 Neural-GPU: a recurrent, local four-row convolutional grid."""
import torch
from torch import nn
from benchmark import OptimizerBundle, Submission, assert_model_state
PAD, N, X, T, ANS, DIGIT = 0, 2, 3, 4, 5, 7
CHANNELS, MAX_PLACES, MAX_STEPS = 48, 16, 64
STATE_ELEMENTS = 126_682
class Config:
def __... | {
"id": "0c81a9a9-f848-4a56-a7e7-f20796628c95",
"created_at": "2026-08-07 14:52:37.347134+00:00",
"db_md5": "d10c959e39f6a42af189c66c5bf39a32",
"submitter": "Ojasw Upadhyay",
"github_login": "oupadhyay",
"run_id": "c5758dea-6132-4d48-9828-9da7f2bfccf1",
"tier": "easy",
"dataset_id": "e5",
"status": "s... | {
"score": {
"mean_loss": 2.1759518597025043,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.005
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.1512075967318274,
"example_count": 600,
"e... |
0c81e8cd-0fc7-429c-a38c-bad139e79452 | easy | viridale | 2026-08-03 12:12:56.513631+00:00 | succeeded | 5f6e63dcae944d9881bfa393d432aa67c964c0501cbe2468308f4ad2e819f75a | 144,755 | null | """Guarded mismatch-triggered three-instruction recurrent tail.
hypothesis: the 27,000 short nested words add useful exact Hard support, while
an explicit empty-index guard makes them cost-free on supported lower paths.
axis: arch
target: revised h1 Max T at least1 with no practice throughput regression.
expected_... | {
"id": "0c81e8cd-0fc7-429c-a38c-bad139e79452",
"created_at": "2026-08-03 12:12:56.513631+00:00",
"db_md5": "b6fd83257eb32b734bcfb475aba2ce5f",
"submitter": "priormancer",
"github_login": "viridale",
"run_id": "82ca920c-c8ea-4c6e-8766-2c64d7788a8c",
"tier": "easy",
"dataset_id": "e5",
"status": "succe... | {
"score": {
"mean_loss": 0.003020370119402878,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 1.0
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 0.003020370181758508,
"example_count": 600,
... |
0c8219d8-3560-4144-83ca-e19c7f173f33 | easy | oupadhyay | 2026-08-25 04:40:28.801961+00:00 | succeeded | b7eab9625894de3d9db82795d0ac7548c1417a194d208e4c26d7e4273888e525 | 4,017 | null | """Shared conditional quadratic recurrence over bidirectional digit states."""
from __future__ import annotations
import torch
import torch.nn.functional as F
from torch import nn
from benchmark import OptimizerBundle,Submission,TokenLossBatch,assert_model_state
W,D,H=4,32,16
class C:
def __init__(s,vocab_size,max_seq... | {
"id": "0c8219d8-3560-4144-83ca-e19c7f173f33",
"created_at": "2026-08-25 04:40:28.801961+00:00",
"db_md5": "5ab2105484d6abc4d29b5d104a200ff9",
"submitter": "Ojasw Upadhyay",
"github_login": "oupadhyay",
"run_id": "26c3c254-36ee-4b62-8c89-358f804d8372",
"tier": "easy",
"dataset_id": "e5",
"status": "s... | {
"score": {
"mean_loss": 2.171819828367153,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.002916666666666667
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.1461369617102926,
"example_count": 600... |
0c8e7be9-6ceb-44f5-b87d-8ff2aa6b1eb8 | easy | yashkant | 2026-08-14 17:27:06.028333+00:00 | succeeded | 681971907ac4328f5f468cd8a9ec2dc2cce26574afa32a2c29f0916dc3e8b132 | 38,959 | null | """Round1818 parent-preserving temporal Anderson-Dykstra mixing.
The complete Round1792 AdamW5/EMA.995 W8/D2 carrier remains byte-identical at
a matched seed. Initialization of the added temporal mixer is isolated from
the inherited CPU RNG stream, and one exact-zero scalar gate keeps the scored
path equal to the par... | {
"id": "0c8e7be9-6ceb-44f5-b87d-8ff2aa6b1eb8",
"created_at": "2026-08-14 17:27:06.028333+00:00",
"db_md5": "67a4412d2f9c24ba228a108a5f2b99e8",
"submitter": "Yash Kant",
"github_login": "yashkant",
"run_id": "5bc9ea27-ce77-491f-ae29-35ccbb5c91e9",
"tier": "easy",
"dataset_id": "e3",
"status": "succeed... | {
"score": {
"mean_loss": 2.1261328010956815,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.01562500015832484
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.1255640363713297,
"example_count": 800... |
0c91c4ef-172b-48e5-8409-bbfe10227391 | easy | 0Chris5R | 2026-08-07 09:35:21.643874+00:00 | succeeded | a783acbef90ff876c736b0d92189b28c16fbbe0549b1b3ba658fc0fd80850d3a | 13,238 | null | """Learned convolutional-GRU over an aligned field-by-digit workspace.
The public field markers and decimal significance define only the layout of
the neural workspace. Every transition and output is learned; there is no
arithmetic routine, recurrence assumption, lookup table, or generated target.
"""
from __future__... | {
"id": "0c91c4ef-172b-48e5-8409-bbfe10227391",
"created_at": "2026-08-07 09:35:21.643874+00:00",
"db_md5": "e3e77101cee08fb266a890e2c0aac977",
"submitter": "Chris ",
"github_login": "0Chris5R",
"run_id": "9dee06ca-589c-4700-ad76-d4edd5f55882",
"tier": "easy",
"dataset_id": "e5",
"status": "succeeded"... | {
"score": {
"mean_loss": 6.737996193075444,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.011666666691501935
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 5.859895971323877,
"example_count": 600,... |
0c966fcc-67b1-4435-91b6-85aa4c66374f | easy | erdavis0 | 2026-08-19 08:52:22.315356+00:00 | succeeded | 578fe3dc387ad5282a7526ec8c3b1ac3e1186867a747f7ee93ff33cfe45f5f79 | 9,293 | null | """Four-step raw-vocabulary operator for shallow learned 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,
TokenLossBa... | {
"id": "0c966fcc-67b1-4435-91b6-85aa4c66374f",
"created_at": "2026-08-19 08:52:22.315356+00:00",
"db_md5": "ead444d24208b7440ac6eda54dbc445a",
"submitter": "Ethan",
"github_login": "erdavis0",
"run_id": "4fe7a3e3-3f04-4afa-a2eb-bd772ed36631",
"tier": "easy",
"dataset_id": "e7",
"status": "succeeded",... | {
"score": {
"mean_loss": 2.136850357055664,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.05213903822004795
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.1373066902160645,
"example_count": 85,
... |
0c9979f8-583e-4eec-ba06-0f7f854ccde2 | easy | arseniivanov | 2026-08-10 02:52:59.934756+00:00 | succeeded | 7bf30e19d90e7f6cadb4ad67d0b2cd51634a4fd2d29a507be6f840a0dd78f8c1 | 3,628 | null | """Basic single-pass Transformer with PyTorch AdamW."""
from __future__ import annotations
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
assert_model_state,
)
D_MODEL = 128
NUM_HEADS = 4
... | {
"id": "0c9979f8-583e-4eec-ba06-0f7f854ccde2",
"created_at": "2026-08-10 02:52:59.934756+00:00",
"db_md5": "eaac69fe345f82973fcf392c504a40ff",
"submitter": "Arseni Ivanov",
"github_login": "arseniivanov",
"run_id": "c00fe8ce-87a3-4e8e-869a-bed5007e7751",
"tier": "easy",
"dataset_id": "e1",
"status": ... | {
"score": {
"mean_loss": 3.0032811164855957,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.013333333656191826
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 3.9463300704956055,
"example_count": 10... |
0c9da856-4819-4dd6-9fd0-777b7a923c46 | easy | arseniivanov | 2026-08-10 02:07:23.339180+00:00 | succeeded | 7d7d7d7060ab864154df7dda05fb53fc28444f1f1acb6069ac9c35d2376c10aa | 18,961 | null | """Looped, vector-quantized, adaptively-halting Transformer.
Depth is created by iteration over a weight-tied step operator, not by stacking layers:
encoder tokens -> N-context untied, read once; hoists the modulus-dependent
work out of the recurrence
st... | {
"id": "0c9da856-4819-4dd6-9fd0-777b7a923c46",
"created_at": "2026-08-10 02:07:23.339180+00:00",
"db_md5": "64f719f8c1526928cb897f7c72b5b6a9",
"submitter": "Arseni Ivanov",
"github_login": "arseniivanov",
"run_id": "8b8d2fe2-2309-494a-8443-502edf321b83",
"tier": "easy",
"dataset_id": "e1",
"status": ... | {
"score": {
"mean_loss": 9.620597839355469,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.04833333194255829
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 10.448322296142578,
"example_count": 100,... |
0ca6e97b-d6f7-400b-bd9e-d90a8b0bc209 | easy | repst | 2026-08-08 12:49:53.972246+00:00 | succeeded | 31ec35ad037dd3cf3741c1f2ccd3782edc99a4069ca81a33a356177877183362 | 5,789 | null | """One-shot relation bottleneck for the T=1 diagnostic.
This model has no recurrent transition or prompt-specific arithmetic. A small
bidirectional reader learns three soft registers, applies one learned bilinear
relation to the value register, and decodes the requested answer digits.
"""
from __future__ import anno... | {
"id": "0ca6e97b-d6f7-400b-bd9e-d90a8b0bc209",
"created_at": "2026-08-08 12:49:53.972246+00:00",
"db_md5": "3c9ae8ab47b1ee1313c412688bdbf868",
"submitter": "Ryan Epstein",
"github_login": "repst",
"run_id": "2b594cd6-cb8f-4c52-81b6-f6de9b350951",
"tier": "easy",
"dataset_id": "e1",
"status": "succeed... | {
"score": {
"mean_loss": 10.348426342010498,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0283333333209157
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 10.607148170471191,
"example_count": 100,... |
0cabd716-3f27-483d-865f-678ebb9f6ece | easy | shan23chen | 2026-08-08 03:13:43.285510+00:00 | succeeded | 547058400e962e0f50990ca266fd6d98ae1e74fed7d63f0fb90e89fcbde4796f | 26,596 | null | """Canonical-encoder recurrent register Transformer.
One learned transition is applied once per requested outer step. The
transition receives an immutable decimal N register and a canonical semantic
digit boundary, but never receives T or a separate copy of the original X.
Within each outer application, scratch persi... | {
"id": "0cabd716-3f27-483d-865f-678ebb9f6ece",
"created_at": "2026-08-08 03:13:43.285510+00:00",
"db_md5": "400a41eb2f88c3c7f1dba1ed9bc800f8",
"submitter": "Shan Chen",
"github_login": "shan23chen",
"run_id": "a10a6c4b-c018-4cd5-9927-5f33357607e7",
"tier": "easy",
"dataset_id": "e5",
"status": "succe... | {
"score": {
"mean_loss": 2.398365020751953,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.007083333446644247
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.2182226181030273,
"example_count": 600... |
0cb09605-c021-4a13-a9a1-55cba79bfcfb | easy | shreyash-chonkie | 2026-09-01 03:36:29.676447+00:00 | succeeded | a31e7251ed1e2910090d4a522bf15f374e699b4d7d9202b356aa481ccebe51da | 20,246 | null | """Learned token scanner feeding a random recurrent path reservoir."""
from __future__ import annotations
from itertools import permutations
import torch
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
TokenLossBatch,
assert_model_s... | {
"id": "0cb09605-c021-4a13-a9a1-55cba79bfcfb",
"created_at": "2026-09-01 03:36:29.676447+00:00",
"db_md5": "1f1fc0968b24c6f2eccbda4e932aef45",
"submitter": "Shreyash",
"github_login": "shreyash-chonkie",
"run_id": "8315ebb0-4880-40da-9133-2c8b32f98886",
"tier": "easy",
"dataset_id": "e9",
"status": "... | {
"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": 90,
"exact_accuracy": 1.0,
"c... |
0cbcf75b-4da0-4fa6-92e7-73d5e5dee69b | easy | yashkant | 2026-08-13 01:06:09.798618+00:00 | succeeded | 394613dfedc841ed31c131359880563f0e7b3e325059ab729a21de0182faa700 | 40,586 | null | """Round1808 fused Dykstra ripple-carry composition.
The exact Round1792 AdamW5 carrier supplies the learned Dykstra primal and
two correction memories. Instead of attaching a separate carry branch, this
candidate right-aligns the raw categorical N/X token embeddings into the
existing Dykstra width and interprets eac... | {
"id": "0cbcf75b-4da0-4fa6-92e7-73d5e5dee69b",
"created_at": "2026-08-13 01:06:09.798618+00:00",
"db_md5": "2faaa19f55a56c2ea8837bfb77b57374",
"submitter": "Yash Kant",
"github_login": "yashkant",
"run_id": "18d8dbe8-594a-4d3e-88f3-ddca2eb5e45e",
"tier": "easy",
"dataset_id": "e3",
"status": "succeed... | {
"score": {
"mean_loss": 2.1224921041363674,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.016875000158324837
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.121948370474148,
"example_count": 800... |
0cc380c8-1787-4093-b513-0090abf07764 | easy | k-penchev | 2026-08-23 00:51:11.041201+00:00 | succeeded | 22036641ab20812af9ac11c50a03d425f58497b8c3be7dfa09addfc6a8fe113c | 9,094 | null | """Schedule-Free recurrence with normalized within-number digit coordinates."""
from __future__ import annotations
from collections.abc import Iterable
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submiss... | {
"id": "0cc380c8-1787-4093-b513-0090abf07764",
"created_at": "2026-08-23 00:51:11.041201+00:00",
"db_md5": "cfa1c816e5fdfdcecf1172cc459ff09b",
"submitter": "Kaloyan Penchev",
"github_login": "k-penchev",
"run_id": "c54c00bf-f153-4444-86f7-90f3a7a239b9",
"tier": "easy",
"dataset_id": "e1",
"status": "... | {
"score": {
"mean_loss": 2.4410025477409363,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0833333320915699
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.8926146030426025,
"example_count": 100,... |
0cc4f41f-ba4a-4275-a0e4-0bd42c54a9dc | easy | karanganesan | 2026-08-05 11:29:44.910881+00:00 | succeeded | 095b1b9ab2008f49b8e29a1953183bbeb985e0efe1bd3b8e5dd40ae38519c4ea | 22,099 | 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": "0cc4f41f-ba4a-4275-a0e4-0bd42c54a9dc",
"created_at": "2026-08-05 11:29:44.910881+00:00",
"db_md5": "6abde6b9be36f19ab5e0caa0415ecfe1",
"submitter": "Karan Ganesan",
"github_login": "karanganesan",
"run_id": "10018e25-6470-455b-b7c4-4c6d2f73ea39",
"tier": "easy",
"dataset_id": "e1",
"status": ... | {
"score": {
"mean_loss": 8.076987743377686,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.011666666716337204
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 8.620450973510742,
"example_count": 100,... |
0cca12ac-6925-4c32-a473-3507a39afcb7 | easy | TZZheng | 2026-08-06 17:46:07.234475+00:00 | succeeded | 4f7a2caff7260de16627a7443566cb0bf27849bdc2a735ffc89847824d8bd4fb | 7,304 | null | """Self-contained standard Transformer scaling-sweep submission."""
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,
)
VARIANT = {'block... | {
"id": "0cca12ac-6925-4c32-a473-3507a39afcb7",
"created_at": "2026-08-06 17:46:07.234475+00:00",
"db_md5": "07b563e92529114fc6ae5f362ff1742c",
"submitter": "TZZheng",
"github_login": "TZZheng",
"run_id": "3e198993-a0ee-4fee-871c-7e085dfd4d89",
"tier": "easy",
"dataset_id": "e1",
"status": "succeeded"... | {
"score": {
"mean_loss": 2.808872163295746,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.06333333067595959
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 3.7099685668945312,
"example_count": 100,... |
0ccaeb99-9035-47db-8d2f-3cf3be37e87a | easy | akkolli | 2026-08-07 02:36:37.407358+00:00 | succeeded | 539b2908b0362f464cbb8e9ded3fc025cc7ee356d724ff73d6fd89784bdbb93f | 203,973 | null | """Ordinal-codebook square lane plus one learned affine-square recurrence.
The ten digit-token categories receive trainable scalar scores under each of
two serialization hypotheses. They start in the participant-visible canonical
token order, which makes fixed-modulus datasets well posed, but remain fully
trainable s... | {
"id": "0ccaeb99-9035-47db-8d2f-3cf3be37e87a",
"created_at": "2026-08-07 02:36:37.407358+00:00",
"db_md5": "6b54184caa101653d0f70478614d4a78",
"submitter": "Akshay Kolli",
"github_login": "akkolli",
"run_id": "44112c34-ec6a-4b44-8cc3-db726df5d3db",
"tier": "easy",
"dataset_id": "e3",
"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": 800,
"exact_accuracy": 1.0,
"... |
0ccb47ac-0c09-4547-abae-c5a2d99194e1 | easy | blake-camp-surge | 2026-08-05 18:47:25.684898+00:00 | succeeded | 37865442a42031df29591c7827b57d90ee65afd2d841c7d08cb4b6cad21de07b | 10,388 | null | """Vector-state neural cellular automaton over attention edges."""
from __future__ import annotations
import math
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
assert_model_state,
)
TOKEN... | {
"id": "0ccb47ac-0c09-4547-abae-c5a2d99194e1",
"created_at": "2026-08-05 18:47:25.684898+00:00",
"db_md5": "a9da34b3cd1a74661d3004d751502444",
"submitter": "Blake Camp",
"github_login": "blake-camp-surge",
"run_id": "c1cd37cf-619c-4776-a3b1-67bd12c8728a",
"tier": "easy",
"dataset_id": "e3",
"status":... | {
"score": {
"mean_loss": 4.115675678863831,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.005625
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 4.202259920502993,
"example_count": 800,
"... |
0ccc0f77-91a2-406d-9196-7f90522f7a29 | easy | yashkant | 2026-08-13 00:41:40.513968+00:00 | succeeded | f7f9386f020f4cbd87f4243a8e51989ec2a344e2fb1b8006dcab536e7615092f | 31,750 | 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": "0ccc0f77-91a2-406d-9196-7f90522f7a29",
"created_at": "2026-08-13 00:41:40.513968+00:00",
"db_md5": "6095e33e176ae791430e4fbd74b9eaa2",
"submitter": "Yash Kant",
"github_login": "yashkant",
"run_id": "df08f345-6a8d-4fd9-af4f-9c97e76fe031",
"tier": "easy",
"dataset_id": "e5",
"status": "succeed... | {
"score": {
"mean_loss": 2.150404545168684,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.011666666679084302
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.1564368602940855,
"example_count": 600... |
0cd2b74e-be19-4f36-9e64-75e5814a1ac7 | easy | khushidahi | 2026-08-04 00:29:31.750441+00:00 | succeeded | 35d9a689a3278bd158f55841133a95f77d61a72ab68c6ba808a801aca95c89cd | 17,188 | 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": "0cd2b74e-be19-4f36-9e64-75e5814a1ac7",
"created_at": "2026-08-04 00:29:31.750441+00:00",
"db_md5": "c857016c92a72edf805e4ec104588745",
"submitter": "khushidahi",
"github_login": "khushidahi",
"run_id": "7a8da6af-51cd-4c9e-995b-6af971b06f68",
"tier": "easy",
"dataset_id": "e4",
"status": "succ... | {
"score": {
"mean_loss": 2.08248507976532,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.00572916679084301
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.080242156982422,
"example_count": 1200,
... |
0cd34bee-019a-4669-8f01-0a2d9ec4c522 | easy | jordanrubin | 2026-08-06 07:54:50.581132+00:00 | succeeded | e56835c78d0796ebd0777f7e1295494210431f22f8d2c60d9b4587b6f39ce8b5 | 22,320 | null | """Dual-channel bidirectional transaction-flow looped Transformer.
The learned outer transition is shared across logical time and T-blind. Its
representation exposes the two opposed information flows required by decimal
quotient/remainder work without installing arithmetic weights or operations:
* low-to-high masked... | {
"id": "0cd34bee-019a-4669-8f01-0a2d9ec4c522",
"created_at": "2026-08-06 07:54:50.581132+00:00",
"db_md5": "658f7700353f8c974f4979d3a3f4936c",
"submitter": "Jordan Rubin",
"github_login": "jordanrubin",
"run_id": "fb46fd4f-1327-46e4-84ce-c9bb1e9802cc",
"tier": "easy",
"dataset_id": "e1",
"status": "s... | {
"score": {
"mean_loss": 2.202615261077881,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.160802125930786,
"example_count": 100,
"exact... |
0cfb6c00-6a5f-459d-9d26-3dda1302b50b | easy | ehonig | 2026-08-20 08:42:46.229920+00:00 | succeeded | bd6c231f9e2ef384d421ff8ae0354dc77963858a0555730ccd8a3da63f43cdb4 | 23,245 | 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": "0cfb6c00-6a5f-459d-9d26-3dda1302b50b",
"created_at": "2026-08-20 08:42:46.229920+00:00",
"db_md5": "302ad6e104ff3a2019d96f2e0e8fd768",
"submitter": "Edouardo Honig",
"github_login": "ehonig",
"run_id": "90cd2286-0b79-48fd-ab39-dd9e3b73ba87",
"tier": "easy",
"dataset_id": "e5",
"status": "succ... | {
"score": {
"mean_loss": 2.2048367501513377,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.006666666610787312
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.2083890438079834,
"example_count": 60... |
0cfc32d1-7b20-4da7-a3e3-f6a433521202 | easy | lzy54 | 2026-08-29 08:24:14.871852+00:00 | succeeded | f967c3bfea2806a17090394364d66f815812d86d38acc90cb7b0b8126f881c01 | 4,539 | null | """Phase 2 C2: shared block x4 with residual concat 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_HEAD... | {
"id": "0cfc32d1-7b20-4da7-a3e3-f6a433521202",
"created_at": "2026-08-29 08:24:14.871852+00:00",
"db_md5": "649789da2b5f7837ff738282292ec92c",
"submitter": "lzy54",
"github_login": "lzy54",
"run_id": "9238be96-710a-459f-ba6e-068f4d797782",
"tier": "easy",
"dataset_id": "e7",
"status": "succeeded",
... | {
"score": {
"mean_loss": 2.9863563776016235,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.03279857523739338
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 3.3072235584259033,
"example_count": 85,... |
0d0390f3-67b3-4764-a7b8-ab7e2897f8a4 | easy | rteehas | 2026-08-19 08:13:50.913352+00:00 | succeeded | 273ad5f793cbe3669c50977e785882ee9367d8cea8806520df9c18be922bb8ea | 41,478 | 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": "0d0390f3-67b3-4764-a7b8-ab7e2897f8a4",
"created_at": "2026-08-19 08:13:50.913352+00:00",
"db_md5": "5679179139776d111a6321f32cbed579",
"submitter": "Ryan Teehan",
"github_login": "rteehas",
"run_id": "494a6b1e-94da-479a-8820-b587c3b82557",
"tier": "easy",
"dataset_id": "e4",
"status": "succee... | {
"score": {
"mean_loss": 7.149502319444824,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0028125
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 7.108493116048063,
"example_count": 1200,
... |
0d0a3465-c508-41dd-a291-a0e0540d3a36 | easy | alirezashirvani-jr | 2026-08-31 00:50:44.877081+00:00 | succeeded | d30c15c4b819074a695cfe7f9ab43fd9530cdf8d2f612f55533983ffcf4e7dfb | 22,445 | 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 = 2
PERIOD_MAXIMUM = 96
EXP... | {
"id": "0d0a3465-c508-41dd-a291-a0e0540d3a36",
"created_at": "2026-08-31 00:50:44.877081+00:00",
"db_md5": "a726edbe0af9de151bd033a13eebcee4",
"submitter": "alirezashirvani-jr",
"github_login": "alirezashirvani-jr",
"run_id": "48093003-def1-4454-870e-869681e3d41d",
"tier": "easy",
"dataset_id": "e8",
... | {
"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": 85,
"exact_accuracy": 1.0,
"c... |
0d0b2229-44f5-4a24-bc44-d13518d6caef | easy | liam-gb | 2026-08-16 16:18:26.928668+00:00 | succeeded | ee8e3602e45d4a044761dc84e5d6b53ee9f63b01800876d8c48a95b4ec2aca98 | 15,754 | null | """rns_pm_gather_ct_widet_ptnore: widet + per-prime temperature (16-prime bank).
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-... | {
"id": "0d0b2229-44f5-4a24-bc44-d13518d6caef",
"created_at": "2026-08-16 16:18:26.928668+00:00",
"db_md5": "5051d831398b6c8922c27ddf8e5e8b36",
"submitter": "liam-gb",
"github_login": "liam-gb",
"run_id": "a36095b3-5a68-420a-a2a3-d81086deeee1",
"tier": "easy",
"dataset_id": "e2",
"status": "succeeded"... | {
"score": {
"mean_loss": 5.099510312080383,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.4133333694189787
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 9.013612747192383,
"example_count": 300,
... |
0d11ef94-f611-4045-90dd-76f2977466d4 | easy | ddoan | 2026-08-14 10:25:08.708345+00:00 | succeeded | 532e3c4207870af672c75e4b179ae8785a294b2515d66b56e8c903068b7b4959 | 7,673 | null | """A compact diagonal Neural GPU active-memory model."""
from __future__ import annotations
import copy
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": "0d11ef94-f611-4045-90dd-76f2977466d4",
"created_at": "2026-08-14 10:25:08.708345+00:00",
"db_md5": "d8e340824273791b2124d3476da51637",
"submitter": "Doug Doan",
"github_login": "ddoan",
"run_id": "95d4daa6-e6e8-4dcb-b0ed-f2bb6adcb59c",
"tier": "easy",
"dataset_id": "e3",
"status": "succeeded"... | {
"score": {
"mean_loss": 2.1489057484827674,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.008750000055879354
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.1521914742501482,
"example_count": 80... |
0d1d7d65-a14e-48f8-a750-5f18eb8db0e8 | easy | isaac0804 | 2026-08-05 08:06:52.973707+00:00 | succeeded | 2d325fb896b4ca833a7cf26aff32f8403b2ad95103f4c947e6768148e5df995a | 11,975 | null | """Weight-tied looped Transformer (v6 core: damped gate, 2 sublayers/iteration)
plus two cheap, legal additions from the standard looped/recurrent-transformer
toolkit, both gated behind module-level flags for fast local ablation:
INPUT INJECTION (``USE_INPUT_INJECTION``): at every loop iteration, re-inject
a lear... | {
"id": "0d1d7d65-a14e-48f8-a750-5f18eb8db0e8",
"created_at": "2026-08-05 08:06:52.973707+00:00",
"db_md5": "e3c154bdc7295c4dd7a5934b95d2e36c",
"submitter": "Isaac Yong",
"github_login": "isaac0804",
"run_id": "ae9a45af-9d3f-4f90-a3cb-d28d57974e84",
"tier": "easy",
"dataset_id": "e1",
"status": "succe... | {
"score": {
"mean_loss": 1.8971669673919678,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.071666669100523
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 1.895523190498352,
"example_count": 100,
... |
0d20fe33-f9ed-4fc7-8652-38e5ea5c664e | easy | velocizapkar | 2026-08-13 10:24:17.585049+00:00 | succeeded | dfda66752c87ce003360c3ff1a2089a6fe6f1c8c0bebdc33dfcfca6802398744 | 10,595 | null | """Geometry-calibrated, completely skipless Transformer with AdamW.
This is the strict L8 calibrated CutSkip 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. Model, initialization,
loss, batch size,... | {
"id": "0d20fe33-f9ed-4fc7-8652-38e5ea5c664e",
"created_at": "2026-08-13 10:24:17.585049+00:00",
"db_md5": "8023d1b9a1aea7f164486c5833bb6fa2",
"submitter": "Aakanksh Zarapkar",
"github_login": "velocizapkar",
"run_id": "bfe2947e-8429-48d7-a6de-028ebdf713c5",
"tier": "easy",
"dataset_id": "e2",
"statu... | {
"score": {
"mean_loss": 2.2706968784332275,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.269514322280884,
"example_count": 300,
"exac... |
0d2a9610-2c54-4480-8ac6-3f5d05aaf9d9 | easy | gauravmishra | 2026-08-10 03:22:04.604386+00:00 | succeeded | 398a8a8a4a29a9bd830ffdb71169f5f6be5bffcad474a9565ca1ea76bc60ccf2 | 7,412 | null | """E07: K13 with matched warmup-stable-decay scheduling."""
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,
)
SCALE ... | {
"id": "0d2a9610-2c54-4480-8ac6-3f5d05aaf9d9",
"created_at": "2026-08-10 03:22:04.604386+00:00",
"db_md5": "2e75180c2f4fac35b5625125d327d76a",
"submitter": "Gaurav Mishra",
"github_login": "gauravmishra",
"run_id": "e97625b1-c0c3-41c9-a0f8-371bd63d1de7",
"tier": "easy",
"dataset_id": "e5",
"status": ... | {
"score": {
"mean_loss": 2.455706758821429,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.005
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.505785579638631,
"example_count": 600,
"exa... |
0d2b74ed-7349-4d56-af09-1d50047e2832 | easy | jayyala | 2026-08-16 22:38:54.920662+00:00 | succeeded | 67093a88f1d7e42eee17a538c42e3d548ccde180b37a92e5bab1d2e0c5e8fb48 | 13,859 | 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": "0d2b74ed-7349-4d56-af09-1d50047e2832",
"created_at": "2026-08-16 22:38:54.920662+00:00",
"db_md5": "67c0ab3c7d809072290bb9f45e40f517",
"submitter": "jayyala",
"github_login": "jayyala",
"run_id": "b17beabf-9cb5-4bf8-bcc1-6d1d553ac5e2",
"tier": "easy",
"dataset_id": "e7",
"status": "succeeded"... | {
"score": {
"mean_loss": 4.633372068405151,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.8641265630722046
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.5365853309631348,
"example_count": 85,
... |
0d317d54-0c6c-40c8-abe2-671fdd90fd24 | easy | yashkant | 2026-08-15 13:09:28.121101+00:00 | succeeded | 5835c5090906a31e0f11494768841784c59e00fe4325e2d747ae6693c783374e | 32,226 | null | """Round1232 parsed-T allocation crossover on the y/z MinMax transition.
The complete Round1220 answer/latent MinMax candidate remains unchanged except
for the public-T-to-macro allocation. Separate answer y and latent z states,
immutable x reinjection, both MinMax updates per macro, and full BPTT remain
intact. MINMA... | {
"id": "0d317d54-0c6c-40c8-abe2-671fdd90fd24",
"created_at": "2026-08-15 13:09:28.121101+00:00",
"db_md5": "25e502d53e86da62dc76f02456a9b310",
"submitter": "Yash Kant",
"github_login": "yashkant",
"run_id": "99bb9df7-7f2b-49b2-a3a0-6fbbed98539b",
"tier": "easy",
"dataset_id": "e5",
"status": "succeed... | {
"score": {
"mean_loss": 2.1458700786906855,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.009583333345750967
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.1490437984466553,
"example_count": 60... |
0d37d55f-ce2e-499f-842a-7b72f13ab452 | easy | pepinu | 2026-08-21 22:54:03.805590+00:00 | succeeded | 613f7882ef36fc8855ca6a5351506070fef00c191bdfe2d75e5077504d9e3e44 | 41,384 | null | """Squaring mod N as a learned permutation on CRT components.
WHY THIS SHAPE. (Z_N, x) is a commutative monoid, so every finite-dimensional
linear representation splits into multiplicative characters and squaring is
diagonal there -- the component basis is the only one in which this operation is
shallow. Two consequen... | {
"id": "0d37d55f-ce2e-499f-842a-7b72f13ab452",
"created_at": "2026-08-21 22:54:03.805590+00:00",
"db_md5": "0cc26802042f1a1f77154f64ae8140f2",
"submitter": "pepinu",
"github_login": "pepinu",
"run_id": "3dfb05af-c186-41fd-a5dd-8479e5099161",
"tier": "easy",
"dataset_id": "e1",
"status": "succeeded",
... | {
"score": {
"mean_loss": 0.014076500199735165,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 1.0
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 0.01408759318292141,
"example_count": 100,
"... |
0d3b9438-3632-49ef-ad94-a95c2ab3ba79 | easy | DDanlov | 2026-08-22 16:09:14.989154+00:00 | succeeded | 2eb3a0021c013fd9c585f8bb65f1f24c39a37e4f00c9d87bf4eff511bc3fb373 | 29,142 | null | """
Unified Prompt-Only Deliberation Model (16M Parameters) - Easy Tier (e5)
Architecture:
- Scale: dim=630, num_heads=21, d_ff=2520 (Exact 15,950,340 parameters ~ 16.0M)
- Deliberation: Prompt-Only Recurrence over [Prompt (L_p), 4 Pause Registers]
- Readout: Cross-Attention Readout Layer predicting full sequence logit... | {
"id": "0d3b9438-3632-49ef-ad94-a95c2ab3ba79",
"created_at": "2026-08-22 16:09:14.989154+00:00",
"db_md5": "7a50b190b718843df9829205f6c8dd31",
"submitter": "DDanlov",
"github_login": "DDanlov",
"run_id": "40fa39f8-4110-45a8-b949-81615d12e771",
"tier": "easy",
"dataset_id": "e5",
"status": "succeeded"... | {
"score": {
"mean_loss": 8.047954442264196,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 8.045883238582867,
"example_count": 600,
"exact... |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.