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 |
|---|---|---|---|---|---|---|---|---|---|---|
1a2bad17-e0ca-4ce8-af1b-6c164e02e433 | easy | erdavis0 | 2026-08-22 07:51:49.320428+00:00 | succeeded | a8c16215bb0f3fa806c67e74285cb0712d9e651566419c00cc7a5c4446638275 | 6,227 | null | """Learned table router with prompt-addressed value banks."""
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 = 64
TABLES = 8
E... | {
"id": "1a2bad17-e0ca-4ce8-af1b-6c164e02e433",
"created_at": "2026-08-22 07:51:49.320428+00:00",
"db_md5": "3a23011babdd74adfa0d1bac8da56c6e",
"submitter": "Ethan",
"github_login": "erdavis0",
"run_id": "95a223d0-e78a-4870-8b6b-14542892ccf5",
"tier": "easy",
"dataset_id": "e8",
"status": "succeeded",... | {
"score": {
"mean_loss": 10.237471580505371,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.046657755970954895
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 9.725834846496582,
"example_count": 85,... |
1a2f4e4f-1f85-4384-9e53-e84358dd9048 | easy | DDanlov | 2026-08-24 03:58:22.008316+00:00 | succeeded | 40e0338451a26aa2278bea0594318d142982b3c7e4f090a9b024d611fc504ff6 | 39,620 | null | """
Unified Hybrid Masked Deliberation Model (44.0M Parameters) - Pure Skipless, No Cross-Attention
Architecture:
- Scale: dim=1536, num_heads=24, d_ff=6144 (Exactly 43,977,216 parameters)
- Hybrid Prefix-Bidirectional + Causal Target Attention Mask
- Token Layout: [Prompt (L_p), 4 Pause Registers, Target Tokens (L_tgt... | {
"id": "1a2f4e4f-1f85-4384-9e53-e84358dd9048",
"created_at": "2026-08-24 03:58:22.008316+00:00",
"db_md5": "7d792fd379ff7153f506fff860526b2d",
"submitter": "DDanlov",
"github_login": "DDanlov",
"run_id": "c42c3906-3b39-4741-b0a8-53845b793b80",
"tier": "easy",
"dataset_id": "e5",
"status": "succeeded"... | {
"score": {
"mean_loss": 2.3762833243010575,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.006666666666666666
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.4148735764849882,
"example_count": 60... |
1a326cec-59a5-458e-a7f7-3eb79afcf69e | easy | DDanlov | 2026-08-13 22:42:24.208045+00:00 | failed | 4b93ad98df76e36861aeb99a43b4d5a81d5730cc162b16e26023f6ca869ff67f | 22,145 | null | import torch
import torch.nn as nn
import torch.nn.functional as F
import math
from dataclasses import dataclass
from typing import Optional
from torch.optim import Optimizer
try:
from client import Submission, assert_model_state, OptimizerBundle
except ImportError:
Submission = None
assert_model_state = N... | {
"id": "1a326cec-59a5-458e-a7f7-3eb79afcf69e",
"created_at": "2026-08-13 22:42:24.208045+00:00",
"db_md5": "5a97f96400669bb63f0c1d27e734611e",
"submitter": "DDanlov",
"github_login": "DDanlov",
"run_id": "5e9f89a5-0746-4be3-8bd4-09acd196ce7e",
"tier": "easy",
"dataset_id": "e5",
"status": "failed",
... | null
|
1a3d8038-b05b-4cca-9709-70d0f565bf18 | easy | velocizapkar | 2026-08-16 06:58:06.715015+00:00 | succeeded | dfbccdf11ac720fa85ab4fbd9c61a231a5d3da8d2776a6f7bd3f6061e6edcddd | 11,023 | null | """Matched standard bidirectional Transformer with deterministic Abacus PE.
This is the ordinary L4 standard-Transformer control with one intervention:
the learned absolute-position table is used instead as a learned decimal-place
table. Each valid digit receives its distance from the right edge of its own
contiguous... | {
"id": "1a3d8038-b05b-4cca-9709-70d0f565bf18",
"created_at": "2026-08-16 06:58:06.715015+00:00",
"db_md5": "642f51af2fe895cc70f1a117b4e5794c",
"submitter": "Aakanksh Zarapkar",
"github_login": "velocizapkar",
"run_id": "0d52c938-4012-4339-8deb-298846ed426e",
"tier": "easy",
"dataset_id": "e6",
"statu... | {
"score": {
"mean_loss": 6.917021036148071,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.02342342445626855
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 7.022780895233154,
"example_count": 60,
... |
1a43e03a-46aa-4ea9-be01-8bf9661c5529 | easy | sapient-sapiens | 2026-08-19 11:17:56.234323+00:00 | succeeded | ecca5854e3a010e2b1475714df197dbbe381639242691b266f81afc1e334dc95 | 32,999 | null | """Depth-quantized recurrence with a mixture-over-depth objective.
The recurrent state is a bank of right-aligned digit slots. Every operator
application is followed by a soft quantization back onto the token simplex, and
the digit logits produced by that quantization are the answer logits at that
depth. There is no s... | {
"id": "1a43e03a-46aa-4ea9-be01-8bf9661c5529",
"created_at": "2026-08-19 11:17:56.234323+00:00",
"db_md5": "ae17018d6e2373d3b67fd34c57483e3e",
"submitter": "Ertondy",
"github_login": "sapient-sapiens",
"run_id": "f32d1087-d3f3-4e2d-89d3-ba5d054f92d9",
"tier": "easy",
"dataset_id": "e9",
"status": "su... | {
"score": {
"mean_loss": 5.484514474868774,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.04555555712431669
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 5.581711292266846,
"example_count": 90,
... |
1a465954-c3c7-49af-8d89-c1e20ae0c8ee | easy | Maxwells-Demons | 2026-08-14 05:22:10.687780+00:00 | succeeded | 39d19782a48458e9b92675994b85fad78096b9058e931d958b4c01aad3f0a243 | 120,067 | null | """One shared digit operator, its two free neighbours of x, and two learned maps.
The backbone is unchanged from v47/v96: the state IS the digit one-hot, one
application of the operator is
(state digits, N digits) -> next state digit logits
and the straight-through code of those logits is the whole state handed to... | {
"id": "1a465954-c3c7-49af-8d89-c1e20ae0c8ee",
"created_at": "2026-08-14 05:22:10.687780+00:00",
"db_md5": "5f70eb0891324e3a78ade8f41ff32bd6",
"submitter": "Maxwells_Ayakashi",
"github_login": "Maxwells-Demons",
"run_id": "801a953c-1073-439d-9f28-bdd9572a879b",
"tier": "easy",
"dataset_id": "e1",
"st... | {
"score": {
"mean_loss": 3.0686267564306036e-05,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 1.0
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.960443453048356e-05,
"example_count": 100,
... |
1a48fc2f-7e44-4cd0-aec2-5e16d6631559 | easy | yashkant | 2026-08-17 01:09:45.414759+00:00 | succeeded | 088b559700b06a4bb0de80361e01c324b21b3e57cac4d5e43a078a8d09e9e55c | 45,397 | null | """Round2219 phase-balanced initialization for an anchored recurrent branch.
The confirmed correction-reset parent receives one width-8 recurrent path.
Its live transition and frozen initial anchor execute in parallel, so their
difference is exactly zero at construction and the parent function is kept.
The candidate s... | {
"id": "1a48fc2f-7e44-4cd0-aec2-5e16d6631559",
"created_at": "2026-08-17 01:09:45.414759+00:00",
"db_md5": "93319034e39dd3769feb83271eac0aa3",
"submitter": "Yash Kant",
"github_login": "yashkant",
"run_id": "94232efe-e64d-4fbe-8c9a-0529ce064368",
"tier": "easy",
"dataset_id": "e5",
"status": "succeed... | {
"score": {
"mean_loss": 2.1493561596803454,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.010833333345750967
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.1541669048001415,
"example_count": 60... |
1a548b0f-ed64-4fae-b861-cfa139b2c6eb | easy | chad-atexpedient | 2026-08-06 03:12:54.287389+00:00 | succeeded | 91d93a7f289df171e56b701bac9479029b01023a68f8335610bd6e6e32145a76 | 10,852 | null | """C02: R028F with AdamW weight decay 0.1."""
from __future__ import annotations
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
assert_model_state,
)
LANES = 6
CHANNELS = 24
MUTABLE_START =... | {
"id": "1a548b0f-ed64-4fae-b861-cfa139b2c6eb",
"created_at": "2026-08-06 03:12:54.287389+00:00",
"db_md5": "f44a5ffc9ec099b945b17c5e026fadf9",
"submitter": "chad-atexpedient",
"github_login": "chad-atexpedient",
"run_id": "c823fe57-ec41-4b11-934f-bd42f3aa758f",
"tier": "easy",
"dataset_id": "e5",
"st... | {
"score": {
"mean_loss": 2.1865367117495795,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.011250000012417634
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.185275808043544,
"example_count": 600... |
1a58c1b2-da57-4afb-b169-036d9ff58532 | easy | adwhit | 2026-08-13 13:33:57.822070+00:00 | succeeded | b300f96288fafe0c84a4a37e2654a54149fc370afe82fe739df9e69db37ec324 | 17,862 | null | """DeepThinker v4a: depth-recurrent transformer with manual CUDA-graph capture.
The evaluation harness is kernel-launch-bound on unrolled-loop models
(measured on H100/Easy: eager 206 steps/60s with the GPU >95% idle;
torch.compile default -> 120 steps net; reduce-overhead -> crash). v4a
captures the ENTIRE training u... | {
"id": "1a58c1b2-da57-4afb-b169-036d9ff58532",
"created_at": "2026-08-13 13:33:57.822070+00:00",
"db_md5": "92e385f3260cf3036ad5e6434ff1a4f5",
"submitter": "adwhit",
"github_login": "adwhit",
"run_id": "c3537226-3977-418a-b66b-bc146da4b7d8",
"tier": "easy",
"dataset_id": "e1",
"status": "succeeded",
... | {
"score": {
"mean_loss": 2.328842282295227,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.03333333320915699
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.6245570182800293,
"example_count": 100,... |
1a5b1be0-2385-4b49-87d1-f98aed01a2fc | easy | sapient-sapiens | 2026-08-19 18:06:19.231983+00:00 | succeeded | 23f40000abf08a3c6dfd43bbea6872f9cee032b39db15bc01720de6ad4457f7f | 34,663 | null | """Depth-quantized recurrence with a mixture-over-depth objective.
The recurrent state is a bank of right-aligned digit slots. Every operator
application is followed by a soft quantization back onto the token simplex, and
the digit logits produced by that quantization are the answer logits at that
depth. There is no s... | {
"id": "1a5b1be0-2385-4b49-87d1-f98aed01a2fc",
"created_at": "2026-08-19 18:06:19.231983+00:00",
"db_md5": "ec8298efa2b0380719cfaca2d079c32b",
"submitter": "Ertondy",
"github_login": "sapient-sapiens",
"run_id": "15e4c72f-dc8c-4960-85af-395f362dc1e3",
"tier": "easy",
"dataset_id": "e5",
"status": "su... | {
"score": {
"mean_loss": 2.183809522423192,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0075
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.183229945032051,
"example_count": 600,
"ex... |
1a5e8388-8655-4c6a-b862-624db0c82840 | easy | jayyala | 2026-08-16 22:37:36.151473+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": "1a5e8388-8655-4c6a-b862-624db0c82840",
"created_at": "2026-08-16 22:37:36.151473+00:00",
"db_md5": "67c0ab3c7d809072290bb9f45e40f517",
"submitter": "jayyala",
"github_login": "jayyala",
"run_id": "78bc6dfd-be38-4d4a-b433-2a4d2966b8d6",
"tier": "easy",
"dataset_id": "e6",
"status": "succeeded"... | {
"score": {
"mean_loss": 1.749722272157669,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.9477477967739105
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 0.5405405163764954,
"example_count": 60,
... |
1a5e8e3b-4fa9-4d4f-b654-68a4e90ee64d | easy | isaac0804 | 2026-08-04 22:01:49.227941+00:00 | succeeded | e150bfc799bbdeb702d126aed06816d9953c9ce4136dea1a005442383096f757 | 8,352 | null | """Weight-tied looped Transformer that reads T from the prompt to set its own depth.
The task is x_t = x_{t-1}^2 mod N for T steps. A single fixed-depth block has no
mechanism to represent "apply this operation T times" and cannot extrapolate to
depths unseen in training. This submission instead applies one shared blo... | {
"id": "1a5e8e3b-4fa9-4d4f-b654-68a4e90ee64d",
"created_at": "2026-08-04 22:01:49.227941+00:00",
"db_md5": "fbe215c54c5741274101702dab493b79",
"submitter": "Isaac Yong",
"github_login": "isaac0804",
"run_id": "1821b29e-17a8-47df-ab08-f0ab232d6a3f",
"tier": "easy",
"dataset_id": "e5",
"status": "succe... | {
"score": {
"mean_loss": 2.211584508329189,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.009583333333333333
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.2305520724822587,
"example_count": 600... |
1a6144c3-2365-4f56-904d-14c3b3947fed | easy | himalalps | 2026-08-07 17:00:28.281153+00:00 | succeeded | cbf0bf1d816ed2a38b916072b5985f4f74a065e7b8b1c3327e45f353fff7f9b9 | 69,661 | null | """Algorithmically constrained latent recurrence for One Layer Deeper.
One masked cross-attention pass creates two independent representations:
``state_0`` from X only and an immutable context from N only. T is excluded
from both representations and is used solely to choose how many times the
learned recurrent MLP tr... | {
"id": "1a6144c3-2365-4f56-904d-14c3b3947fed",
"created_at": "2026-08-07 17:00:28.281153+00:00",
"db_md5": "7418f51d2372b9538294ea072a6402d0",
"submitter": "Haoyu Tang",
"github_login": "himalalps",
"run_id": "11f4ee2c-2bc6-4e02-a39e-9c4acb26a3d9",
"tier": "easy",
"dataset_id": "e5",
"status": "succe... | {
"score": {
"mean_loss": 3.5951712805408613,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.00833333342646559
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.332833766937256,
"example_count": 600,... |
1a6a6ac5-2b8c-43b4-bbc3-ebc18689788b | easy | viridale | 2026-08-03 22:51:01.976933+00:00 | succeeded | 2394f4270c2c90255fa93d69719c0f052fdcbb973bb18204abf1132dd6ef4f80 | 147,614 | null | """Mismatch-triggered maximal five-instruction recurrent-program tail.
hypothesis: the hidden recurrence may be a short exact program of length five;
this is the deepest complete 30-primitive word language below the ceiling.
axis: arch
target: revised h1 Max T at least1 while preserving every v149 learned value.
e... | {
"id": "1a6a6ac5-2b8c-43b4-bbc3-ebc18689788b",
"created_at": "2026-08-03 22:51:01.976933+00:00",
"db_md5": "5f506dfe69578f373a14e9f7cd430d78",
"submitter": "priormancer",
"github_login": "viridale",
"run_id": "b2a6d5ce-297e-446a-a5ea-a07d4aba3cbc",
"tier": "easy",
"dataset_id": "e4",
"status": "succe... | {
"score": {
"mean_loss": 0.003020250739847862,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 1.0
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 0.003020250787608155,
"example_count": 1200,
... |
1a717b78-aedf-49d2-a171-c18d317a8927 | easy | ddoan | 2026-08-14 09:20:50.357545+00:00 | succeeded | 011ede04fc25f9961c7d6d72830324ae3e539a45f55a7bfa45a9cfc21e2be71a | 9,572 | null | """Batch-600 recurrence with learned anti-diagonal pair convolution."""
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": "1a717b78-aedf-49d2-a171-c18d317a8927",
"created_at": "2026-08-14 09:20:50.357545+00:00",
"db_md5": "fa0952ae3388518b08df35f4e66cb31d",
"submitter": "Doug Doan",
"github_login": "ddoan",
"run_id": "747f1595-14d6-478b-bac0-aceb33a3bdef",
"tier": "easy",
"dataset_id": "e4",
"status": "succeeded"... | {
"score": {
"mean_loss": 2.0890488681284975,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.004791666666666666
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.0951361724495237,
"example_count": 12... |
1a74797d-f741-4ed2-bc63-165687393a37 | easy | chad-atexpedient | 2026-08-08 06:49:06.722194+00:00 | succeeded | d97fe4e83bb862a454d58b49bfac3de789da2f0a97672554c2f104910328ab7e | 17,155 | null | """S02: S01 with scratch-only modulus digit reconstruction."""
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,
)
WI... | {
"id": "1a74797d-f741-4ed2-bc63-165687393a37",
"created_at": "2026-08-08 06:49:06.722194+00:00",
"db_md5": "5215a4a164ffe2b68ad760a6f3fbaad2",
"submitter": "chad-atexpedient",
"github_login": "chad-atexpedient",
"run_id": "fa3ee3ff-7a09-42de-b1eb-0be532b15ee5",
"tier": "easy",
"dataset_id": "e5",
"st... | {
"score": {
"mean_loss": 2.3656492524856745,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.010416666691501935
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.43556398447319,
"example_count": 600,... |
1a75f618-11a6-44a7-b329-a6156c605d6c | easy | DDanlov | 2026-08-08 11:40:36.485012+00:00 | succeeded | 8d64595f37de5418ae8cdf2740637089d31382d2a503c021f510c2937c760fa4 | 16,060 | null | from __future__ import annotations
import sys
import os
import math
import time
import torch
import torch.nn as nn
import torch.nn.functional as F
from torch.optim.optimizer import Optimizer
VOCAB_SIZE = 17
DIGIT_OFFSET = 7
class RohanShampoo(Optimizer):
"""Self-contained RohanShampoo optimizer with eigenvalue ma... | {
"id": "1a75f618-11a6-44a7-b329-a6156c605d6c",
"created_at": "2026-08-08 11:40:36.485012+00:00",
"db_md5": "0d9ba3a1e484109a89a4e34ef410346b",
"submitter": "DDanlov",
"github_login": "DDanlov",
"run_id": "8a98fcdf-fdc3-4872-87d5-cb93c7966b54",
"tier": "easy",
"dataset_id": "e1",
"status": "succeeded"... | {
"score": {
"mean_loss": 25.178351659132655,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0533333332836628
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 31.37761688232422,
"example_count": 100,
... |
1a7da1e7-d036-40e8-9135-e1c019d811b7 | easy | YangWang92 | 2026-08-17 05:39:33.705972+00:00 | succeeded | 44fc930d809c31a6480ef79b59b1620f4f36943377541193bab7b764615787de | 46,820 | null | """Limb (bignum) circuit submission.
The h1 metrics pinned the hard tier at ~60-64 bit moduli (the training-loss
floor of 100-250 is exactly the fp64 quantisation error of 18-19 digit
integers), so every value here is a little-endian limb vector, base 10^4.
No intermediate ever exceeds ~1.4e9, so a single float64 hold... | {
"id": "1a7da1e7-d036-40e8-9135-e1c019d811b7",
"created_at": "2026-08-17 05:39:33.705972+00:00",
"db_md5": "7c68cdea41972b37f3eb5b23a332c615",
"submitter": "Yang Wang ",
"github_login": "YangWang92",
"run_id": "095fd84e-b660-427b-8068-0f1a5dcc2e43",
"tier": "easy",
"dataset_id": "e4",
"status": "succ... | {
"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": 1200,
"exact_accuracy": 1.0,
... |
1a7f9cdf-01d5-4972-85ec-1e7647faab8c | easy | blake-camp-surge | 2026-08-04 16:57:35.050631+00:00 | succeeded | 6fc06e7335c2584894a47aa72f4f9ca35bd361647d15b84f520d06c144bd6da8 | 14,288 | null | """Recurrent Transformer with residual DAN-generated hyper-attention."""
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,
ass... | {
"id": "1a7f9cdf-01d5-4972-85ec-1e7647faab8c",
"created_at": "2026-08-04 16:57:35.050631+00:00",
"db_md5": "e7d5b392ebe413078e612ccddacbdf1c",
"submitter": "Blake Camp",
"github_login": "blake-camp-surge",
"run_id": "15f1e046-05c9-4650-84b3-bbd80ea29839",
"tier": "easy",
"dataset_id": "e3",
"status":... | {
"score": {
"mean_loss": 4.8403221422649825,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.008125000009313226
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 4.934915590346436,
"example_count": 800... |
1a832e77-3290-4d1c-8cff-70f8df84a790 | easy | Alan-Qin | 2026-08-18 17:16:06.594337+00:00 | succeeded | 0bc5d50fb167839ac0b3414e658dfef674352a761e2d3917c11b42c088a3bbec | 16,654 | null | """One-hot CRT loop + sibling-consistency supervision (v24).
Architecture: per-band one-hot residues (parameter-free encoding of x),
a learned per-band transition matrix W_m applied T times (STE one-hot), and a
shared digit decoder read at every loop step.
Training loss (token_training_loss): besides the final-answer... | {
"id": "1a832e77-3290-4d1c-8cff-70f8df84a790",
"created_at": "2026-08-18 17:16:06.594337+00:00",
"db_md5": "cf3600bf05b023d27d45e55095750f32",
"submitter": "Zeyu Qin",
"github_login": "Alan-Qin",
"run_id": "9efd79c5-7784-4e82-9512-39058406849b",
"tier": "easy",
"dataset_id": "e7",
"status": "succeede... | {
"score": {
"mean_loss": 0.27597618103027344,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.9461675882339478
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 0.29089874029159546,
"example_count": 85... |
1a8df5bc-2578-43d3-91ee-5889798fabee | easy | jordanrubin | 2026-08-14 01:31:18.903969+00:00 | succeeded | 69ece06bd94846b5a380dc94d4c00c6798b85f8c1e9f0ba91ba31bcdb9376a77 | 29,580 | 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": "1a8df5bc-2578-43d3-91ee-5889798fabee",
"created_at": "2026-08-14 01:31:18.903969+00:00",
"db_md5": "0cdede812da1d9957f86c9ea89ed1f0f",
"submitter": "Jordan Rubin",
"github_login": "jordanrubin",
"run_id": "67a57095-13b8-48a5-9acb-7d3b4a5918b1",
"tier": "easy",
"dataset_id": "e5",
"status": "s... | {
"score": {
"mean_loss": 2.50994313303233,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.08583333211330076
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.1472862525905727,
"example_count": 600,
... |
1a8f9f8c-26d4-4b2d-a9c3-fb3b04a484d5 | easy | DDanlov | 2026-08-18 10:46:38.416857+00:00 | succeeded | 9e2e886a7f72218c6f6788250c79c4edcba3908cfb09a02021dbdad2941008ba | 23,584 | null | """
Official Submission for One Layer Deeper Challenge
Architecture: 4-Layer Recurrent Core Pure No-Norm No-Skip DKS Reasoning Model:
- 4 Distinct Pure No-Norm No-Skip Transformer Layers in the Recurrent Core: L1 -> L2 -> L3 -> L4
- Unrolled T times for serial modular squaring
- dim=512, 16 heads, d_ff=1024, RohanShamp... | {
"id": "1a8f9f8c-26d4-4b2d-a9c3-fb3b04a484d5",
"created_at": "2026-08-18 10:46:38.416857+00:00",
"db_md5": "d6174d2a9211082098f18de894eca987",
"submitter": "DDanlov",
"github_login": "DDanlov",
"run_id": "529e34e4-488c-4d73-9869-147a5b89a692",
"tier": "easy",
"dataset_id": "e1",
"status": "succeeded"... | {
"score": {
"mean_loss": 2.038943359798105,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.06500000054637592
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.0066845657220527,
"example_count": 100,... |
1a93ec2a-6acd-4005-83b9-fa974d7ab350 | easy | DDanlov | 2026-08-24 04:26:26.489906+00:00 | succeeded | 602f787f39d2be99a636634805267e3d4242a4422df47f33a13ea12685fdd6ff | 40,842 | null | """
Unified Hybrid Masked Deliberation Model (44.0M Parameters) - Pure Skipless, No Cross-Attention
Architecture:
- Scale: dim=1536, num_heads=24, d_ff=6144 (Exactly 43,978,752 parameters)
- Hybrid Prefix-Bidirectional + Causal Target Attention Mask
- Token Layout: [Prompt (L_p), 4 Pause Registers, Target Tokens (L_tgt... | {
"id": "1a93ec2a-6acd-4005-83b9-fa974d7ab350",
"created_at": "2026-08-24 04:26:26.489906+00:00",
"db_md5": "0b60d50656733e028b5afb710ef10fb0",
"submitter": "DDanlov",
"github_login": "DDanlov",
"run_id": "8acde407-1e94-4540-bf0b-0455ed53412e",
"tier": "easy",
"dataset_id": "e5",
"status": "succeeded"... | {
"score": {
"mean_loss": 3.36286792961629,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 3.378856224864053,
"example_count": 600,
"exact_... |
1a9c2247-f8bb-49f4-ada4-edc27f6c60fc | easy | Koenneker | 2026-08-05 09:15:31.551338+00:00 | succeeded | 782cabac62246cfe9ac2fa798be45be80cc123cc3ab8298fe3d528cc9b196bab | 2,768 | null | """Vanilla single-layer LSTM with PyTorch AdamW."""
from __future__ import annotations
import torch
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
assert_model_state,
)
D_MODEL = 128
HIDDEN_SIZE = 256
NUM_LAYERS = 1
class Config:
... | {
"id": "1a9c2247-f8bb-49f4-ada4-edc27f6c60fc",
"created_at": "2026-08-05 09:15:31.551338+00:00",
"db_md5": "49ac6a3b8034b9b4b7fd8946e42672a8",
"submitter": "Ferdinand Könneker",
"github_login": "Koenneker",
"run_id": "503e282b-ee6d-4ca9-af1a-6f1b5cd83f00",
"tier": "easy",
"dataset_id": "e3",
"status"... | {
"score": {
"mean_loss": 8.590139029012168,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.004375000009313226
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 7.896504243281793,
"example_count": 800,... |
1a9d9ed4-88d1-4f8e-a0c2-caa3a6d66267 | easy | ehonig | 2026-08-21 01:35:06.617777+00:00 | succeeded | a8101c7bf20f5fb52423225be6359acd59e36c1d9a8a08942c855acf40ae70ac | 31,882 | 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": "1a9d9ed4-88d1-4f8e-a0c2-caa3a6d66267",
"created_at": "2026-08-21 01:35:06.617777+00:00",
"db_md5": "75d8fa2efbdc026e18b4eec8692e50b8",
"submitter": "Edouardo Honig",
"github_login": "ehonig",
"run_id": "a7e5a10e-79dc-4b2d-b846-de338aa208be",
"tier": "easy",
"dataset_id": "e4",
"status": "succ... | {
"score": {
"mean_loss": 2.2226487838136935,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.005520833345750968
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.228111899071329,
"example_count": 120... |
1a9e7593-e261-4399-b4b7-9f8999103d9b | easy | 0Chris5R | 2026-08-01 07:10:51.281340+00:00 | succeeded | d27d56ad9d4881dd4c969482466093710aed3fce8b665221209db8791db3e7fe | 20,798 | null | """Identity-initialized, input-recalling T-REX Neural-GPU.
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 two CGRU layers ... | {
"id": "1a9e7593-e261-4399-b4b7-9f8999103d9b",
"created_at": "2026-08-01 07:10:51.281340+00:00",
"db_md5": "a8757fdfd5cfb3d3daa381d4e28d04b4",
"submitter": "Chris ",
"github_login": "0Chris5R",
"run_id": "09b7c839-3b3a-4290-96be-860a034b8df7",
"tier": "easy",
"dataset_id": "e5",
"status": "succeeded"... | {
"score": {
"mean_loss": 4.923212500280579,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.014583333345750966
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 4.899349819918919,
"example_count": 600,... |
1a9fcbb7-ba0e-4b7b-b1cf-b326fc1aaf58 | easy | repst | 2026-08-24 02:58:34.935174+00:00 | succeeded | 7bcef057d7a4eba0aae7e1940fb0ac08a624476a963f0878588723a3cb5fb90d | 23,255 | null | # R569 DIAGNOSTIC: R568 with fixed-confidence evaluation logits.
#
# This is a measurement-only revision. Training and every evaluation
# argmax are unchanged; fixed confidence makes hosted token loss reveal
# per-digit accuracy. It claims no predictive improvement.
#
# This standalone source extracts the active archit... | {
"id": "1a9fcbb7-ba0e-4b7b-b1cf-b326fc1aaf58",
"created_at": "2026-08-24 02:58:34.935174+00:00",
"db_md5": "d27832cdc5b62340fe46f74d01107adb",
"submitter": "Ryan Epstein",
"github_login": "repst",
"run_id": "35786576-60e3-4a6e-b3dc-272036221987",
"tier": "easy",
"dataset_id": "e1",
"status": "succeed... | {
"score": {
"mean_loss": 0.2231435477733612,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 1.0
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 0.2231435477733612,
"example_count": 100,
"exa... |
1aae8e72-7bb2-4c07-90dc-dc65b8ad2660 | easy | heathsanchez | 2026-08-17 01:51:46.522302+00:00 | succeeded | 1ebd3b7a6298ef414caab2c5c29aff8f4d3359824697773a14f46cdaa297cecd | 3,226 | null | """Frozen official One Layer Deeper baseline (copied from upstream)."""
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
class Config:... | {
"id": "1aae8e72-7bb2-4c07-90dc-dc65b8ad2660",
"created_at": "2026-08-17 01:51:46.522302+00:00",
"db_md5": "a192c4e7033796f4bb288afb6dc6eabb",
"submitter": "Heath Sanchez",
"github_login": "heathsanchez",
"run_id": "5948ea1f-2798-4811-9cce-c381c679fc34",
"tier": "easy",
"dataset_id": "e1",
"status": ... | {
"score": {
"mean_loss": 3.0222411155700684,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.013333333656191826
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 3.955763101577759,
"example_count": 100... |
1ab973ee-cfad-4218-8c9e-7e094334d01b | easy | khushidahi | 2026-08-16 18:52:24.623029+00:00 | succeeded | 058cb022ec54f48bd7a1c64a85f27d87d886f67a3c2df6bb773d56a11e99d35b | 24,892 | null | """R158: structurally symmetric learned digit-pair transition.
This candidate is deliberately diagnostic. It trains on evaluator-owned final
labels at several depths so that one shared transition must compose. The model
never converts N or X to native numbers and never implements multiplication or
modular reduction.... | {
"id": "1ab973ee-cfad-4218-8c9e-7e094334d01b",
"created_at": "2026-08-16 18:52:24.623029+00:00",
"db_md5": "d9fa800598c693228fcf41a50f7f2301",
"submitter": "khushidahi",
"github_login": "khushidahi",
"run_id": "a2a58a3c-e538-446d-a417-3e2cc7ff14a0",
"tier": "easy",
"dataset_id": "e10",
"status": "suc... | {
"score": {
"mean_loss": 1.3821995258331299,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.6392525434494019
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 1.1225663423538208,
"example_count": 125,... |
1abecfd2-5e46-4fb9-ba20-34db95f33664 | easy | jordanrubin | 2026-08-05 21:13:39.436494+00:00 | succeeded | 61d01958ddc3afd2fdccd8de850ed03f3c6dc3226798b31cf9bd59fc3fbacb10 | 24,859 | null | """Adaptive exploration template for the exact-T looped Transformer.
The T=1 computation is the proven value-Fourier baseline cell: a D=256,
three-block bidirectional Transformer with abacus digit positions and learned
Fourier features over the parsed integer x. For T>1 the *same* three-block
cell is reused exactly T... | {
"id": "1abecfd2-5e46-4fb9-ba20-34db95f33664",
"created_at": "2026-08-05 21:13:39.436494+00:00",
"db_md5": "605d906b1c7e83016c01e78f341eca75",
"submitter": "Jordan Rubin",
"github_login": "jordanrubin",
"run_id": "fe30f190-0032-4639-ac1f-93b52796ac02",
"tier": "easy",
"dataset_id": "e3",
"status": "s... | {
"score": {
"mean_loss": 2.476469045843341,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.007499999832361937
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.4558031722483236,
"example_count": 800... |
1ac0e710-5021-4ff0-a8b9-c24a1f7a7ba7 | easy | isaac0804 | 2026-08-04 16:37:29.183302+00:00 | succeeded | 946161a2dd374e652b979936e9f9aba3d5cb0b8b1b6ae58f77775e2c5249978f | 6,910 | null | """Weight-tied looped Transformer that reads T from the prompt to set its own depth.
The task is x_t = x_{t-1}^2 mod N for T steps. A single fixed-depth block has no
mechanism to represent "apply this operation T times" and cannot extrapolate to
depths unseen in training. This submission instead applies one shared blo... | {
"id": "1ac0e710-5021-4ff0-a8b9-c24a1f7a7ba7",
"created_at": "2026-08-04 16:37:29.183302+00:00",
"db_md5": "d7a688e1999e1a65266208ed0c343b81",
"submitter": "Isaac Yong",
"github_login": "isaac0804",
"run_id": "4944a4d5-dff6-491a-a533-4c3d653dd5e4",
"tier": "easy",
"dataset_id": "e1",
"status": "succe... | {
"score": {
"mean_loss": 2.0302376747131348,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.0818095207214355,
"example_count": 100,
"exa... |
1ac0f695-302d-4558-aa26-713daa36a989 | easy | chuk-uzowihe | 2026-08-09 20:56:28.420370+00:00 | succeeded | 8daea7106dd7a8a31fe91fcd458ccf212be7877c462f209f586df2cd0dfab143 | 39,558 | null | """One Layer Deeper submission: depth-routed MoEUT over a 2D digit lattice.
v9: the recurrent block is a single MoEUT layer (refs/moeut) — SwitchHead
attention + sigma-MoE feedforward — whose routers are conditioned on the
loop depth (Fourier features of the iteration index) so one shared bank can
express phase-depend... | {
"id": "1ac0f695-302d-4558-aa26-713daa36a989",
"created_at": "2026-08-09 20:56:28.420370+00:00",
"db_md5": "c7fa80983103a4f807b96ac98cab168f",
"submitter": "Chuk Uzowihe",
"github_login": "chuk-uzowihe",
"run_id": "c5b28019-75f6-4473-af70-628a016979b0",
"tier": "easy",
"dataset_id": "e1",
"status": "... | {
"score": {
"mean_loss": 6.772323846817017,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.011666666716337204
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 6.677601337432861,
"example_count": 100,... |
1ac465de-1aec-422a-9110-68c6d63556e4 | easy | Laz4rz | 2026-08-08 04:24:43.880989+00:00 | succeeded | feef559a7d661a0cf32a91a9bf378ce45534c3a8391c8b2c25b36f4b3606e066 | 36,830 | null | """Standalone diffusion-NextLat DLMS submission entrypoint.
The primary sampler is exactly plain uniform-p1 DLMS. During training, an
adjacent target state supplies detached latent and decoder-distribution targets
to a small pointwise residual dynamics head.
"""
from __future__ import annotations
import math
from ty... | {
"id": "1ac465de-1aec-422a-9110-68c6d63556e4",
"created_at": "2026-08-08 04:24:43.880989+00:00",
"db_md5": "013df82c328ec0c3134be41e729e1b7b",
"submitter": "Laz4rz",
"github_login": "Laz4rz",
"run_id": "a1e8742c-6c6e-4f9a-b6d1-9e548c25a537",
"tier": "easy",
"dataset_id": "e1",
"status": "succeeded",
... | {
"score": {
"mean_loss": 2.751874327659607,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.03333333320915699
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.7345316410064697,
"example_count": 100,... |
1ac7a6e1-fa8a-4a78-b4a2-f3fa0f81dbd7 | easy | sapient-sapiens | 2026-08-15 20:55:40.821649+00:00 | succeeded | ae1d9d9680e9968770f728476fb2063da8af83fae2dc6598fdb449964ecd9f7d | 21,402 | 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": "1ac7a6e1-fa8a-4a78-b4a2-f3fa0f81dbd7",
"created_at": "2026-08-15 20:55:40.821649+00:00",
"db_md5": "fa40672ef26b85fa1966250eaa2856d4",
"submitter": "Ertondy",
"github_login": "sapient-sapiens",
"run_id": "1af093b8-64be-42aa-94e5-b146df97551e",
"tier": "easy",
"dataset_id": "e7",
"status": "su... | {
"score": {
"mean_loss": 6.040049893515451,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.02860962578349493
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 6.211194038391113,
"example_count": 85,
... |
1ac980a6-f13d-4463-85b0-0067c21891e5 | easy | rteehas | 2026-08-19 08:24:20.312609+00:00 | succeeded | 7c89da12c3e257a53f9f75a86eaf44cab870da3af4f027de65b2b1a6259721d3 | 41,409 | 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": "1ac980a6-f13d-4463-85b0-0067c21891e5",
"created_at": "2026-08-19 08:24:20.312609+00:00",
"db_md5": "09ac2233c31b4fc6e188df5974579674",
"submitter": "Ryan Teehan",
"github_login": "rteehas",
"run_id": "758615c1-47bd-487c-b118-b4c7431fbf44",
"tier": "easy",
"dataset_id": "e4",
"status": "succee... | {
"score": {
"mean_loss": 6.869793568121466,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.004895833333333334
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 6.84740620022932,
"example_count": 1200,... |
1ad962fe-9cc3-4c4b-9919-1d10da53e8a9 | easy | richardcepka | 2026-08-23 17:26:35.083477+00:00 | failed | 6b2ef54701284dd2a170d566b9e09b5270006d4581bdd08333468fe0ff1d3c82 | 44,425 | null | """Recurrent digit-register model for One Layer Deeper.
Given a modulus N, a starting value x, and a step count T, predict the residue after squaring modulo N exactly T times.
repo/README.md
A looped transformer whose recurrent state is a decimal digit register. The loop is nested: an
outer step is one squaring, an... | {
"id": "1ad962fe-9cc3-4c4b-9919-1d10da53e8a9",
"created_at": "2026-08-23 17:26:35.083477+00:00",
"db_md5": "edf5092a761357998957395ef3002719",
"submitter": "Richard Cepka",
"github_login": "richardcepka",
"run_id": "27cb778a-c32c-4c5e-af7f-f3ebdbfe2bbe",
"tier": "easy",
"dataset_id": "e1",
"status": ... | null
|
1ae072a6-7e34-464b-b193-be2c9dacb6a9 | easy | shreyash-chonkie | 2026-08-26 18:01:16.263297+00:00 | succeeded | 2468677e32c1e002c6135f5e8413570bf43c895026bde9f29a87b636dea6f91e | 17,833 | null | """Online hard controller over rank-one associative memory updates."""
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,
)
P... | {
"id": "1ae072a6-7e34-464b-b193-be2c9dacb6a9",
"created_at": "2026-08-26 18:01:16.263297+00:00",
"db_md5": "7509b2b3f470bbd5cdf2cc2a5c344d64",
"submitter": "Shreyash",
"github_login": "shreyash-chonkie",
"run_id": "319f14cf-1679-4631-bee8-8eb208044be8",
"tier": "easy",
"dataset_id": "e1",
"status": "... | {
"score": {
"mean_loss": 1.9648644924163818,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.053333332762122154
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 1.9293217658996582,
"example_count": 10... |
1ae2719f-c8c9-44ee-ad81-dd2d3a0a0c82 | easy | chuk-uzowihe | 2026-08-12 19:12:18.422707+00:00 | succeeded | 42641b25e01b384b4e7b29d190b08e0b786e5c68708a9954343dadaa6f25d8dc | 26,028 | null | """v10: dense fixed-loop recurrence with deep supervision.
Lineage: FPRM structure (alpha/beta-tied residual scaling, pre-rms_norm,
input injection every iteration) + the v8 2D digit lattice / RoPE task
encoding. Three deliberate departures from the v9 MoEUT line:
1. **Full BPTT.** The loop runs a fixed N_ITER iterat... | {
"id": "1ae2719f-c8c9-44ee-ad81-dd2d3a0a0c82",
"created_at": "2026-08-12 19:12:18.422707+00:00",
"db_md5": "6c4ac5e64254a081c410d894b8bdea9b",
"submitter": "Chuk Uzowihe",
"github_login": "chuk-uzowihe",
"run_id": "977dfbb2-2a7d-4704-86fd-5c1e176ed4a3",
"tier": "easy",
"dataset_id": "e1",
"status": "... | {
"score": {
"mean_loss": 4.256220579147339,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.06333333253860474
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 3.547236442565918,
"example_count": 100,
... |
1ae66255-4191-4b1c-9650-86886bcee40c | easy | sapient-sapiens | 2026-08-23 08:58:14.683643+00:00 | succeeded | 9d4f52353d497591781716a3ec3f72e22c48f276f4ce1b28fc6bd402f1f2d19e | 20,369 | null | """Approximately 10x-scale adaptive child of the D6 pair/quotient architecture.
The architecture uses the required learned additive token/group/left-relative
representation. Marker-derived field lengths additionally route categorical
digit embeddings into right-aligned lanes; this alternative place alignment is
an ex... | {
"id": "1ae66255-4191-4b1c-9650-86886bcee40c",
"created_at": "2026-08-23 08:58:14.683643+00:00",
"db_md5": "68037a9702f80e9dc822c0ec6cb7ec2e",
"submitter": "Ertondy",
"github_login": "sapient-sapiens",
"run_id": "51637d0e-5c00-4d00-b458-43413de750f8",
"tier": "easy",
"dataset_id": "e5",
"status": "su... | {
"score": {
"mean_loss": 2.1922631756269473,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.006250000024835269
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.1959132018912535,
"example_count": 60... |
1ae70069-6eaf-4042-9060-b6359a332deb | easy | 0xDEADD06 | 2026-08-12 13:42:25.102778+00:00 | succeeded | 21f35c56424581175de95e6fd17cfc70d242e85ad29e1c2e9460e7f923b9f75a | 3,979 | null | """CNN: 1D convolutional stack for algorithmic tasks."""
from __future__ import annotations
import torch
import torch.nn.functional as F
from torch import nn, Tensor
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
assert_model_state,
)
D_MODEL = 256
NUM_LAYERS = 4... | {
"id": "1ae70069-6eaf-4042-9060-b6359a332deb",
"created_at": "2026-08-12 13:42:25.102778+00:00",
"db_md5": "3516eb4dd0f7d32c6fae376597e4e1ab",
"submitter": "0xDEADD06",
"github_login": "0xDEADD06",
"run_id": "df476019-5bff-49da-803f-df61a5f4c907",
"tier": "easy",
"dataset_id": "e5",
"status": "succee... | {
"score": {
"mean_loss": 3.1346812491163214,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.006666666846722364
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 3.0433793238994786,
"example_count": 60... |
1ae88ff0-2802-45fc-936b-c9e68852c50d | easy | oupadhyay | 2026-08-24 01:29:36.901772+00:00 | succeeded | a4174e41706adb433ef1d93c3c45e40b0e3cc6bac0f36646b85d0b711997ceea | 4,247 | null | """Pairwise digit-relation scan for the Easy tier."""
from __future__ import annotations
import torch
import torch.nn.functional as F
from torch import nn
from benchmark import ModelSpec,OptimizerSpec,OptimizerBundle,Submission,TokenLossBatch,assert_model_state
W,D,H=4,48,96
class C:
def __init__(s,vocab_size,max_seq_... | {
"id": "1ae88ff0-2802-45fc-936b-c9e68852c50d",
"created_at": "2026-08-24 01:29:36.901772+00:00",
"db_md5": "b5371a0cfd403cfee71501667de01494",
"submitter": "Ojasw Upadhyay",
"github_login": "oupadhyay",
"run_id": "7e88f771-41fc-4ca9-b478-fb3d72e12550",
"tier": "easy",
"dataset_id": "e5",
"status": "s... | {
"score": {
"mean_loss": 2.2023474293436367,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.00791666670391957
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.160978410276062,
"example_count": 600,... |
1ae98c5a-1d71-49b2-96ac-608823e919e7 | easy | jordanrubin | 2026-08-05 19:25:43.387179+00:00 | succeeded | ab209c56a8280eb97efab73f0873e6aa0f10f9c99814daa1384bed6f3466689a | 9,435 | null | """L1 continuous loop with a learned identity-biased transition step.
The ordinary one-block update is interpreted as a proposed endpoint and a
learned per-channel gate chooses the recurrent Euler step toward it. Gates are
randomly initialized near 0.25, preserving one-block dose while biasing the
shared map toward s... | {
"id": "1ae98c5a-1d71-49b2-96ac-608823e919e7",
"created_at": "2026-08-05 19:25:43.387179+00:00",
"db_md5": "789ce0e7d0194526b87cb4f6e28032bb",
"submitter": "Jordan Rubin",
"github_login": "jordanrubin",
"run_id": "3c4386bb-8e0a-4cf3-8d16-0b31811299d4",
"tier": "easy",
"dataset_id": "e2",
"status": "s... | {
"score": {
"mean_loss": 5.067484282519469,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.05874999934652199
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 6.75334088782954,
"example_count": 300,
... |
1aee435e-864f-43c4-b932-1075d40886d7 | easy | karanganesan | 2026-08-06 05:52:11.101656+00:00 | succeeded | 589b38f5d8707aebc37328dca8bb3402d585d9879dabd0b381e51c09cac94312 | 25,583 | 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": "1aee435e-864f-43c4-b932-1075d40886d7",
"created_at": "2026-08-06 05:52:11.101656+00:00",
"db_md5": "103973e92ee5950bafbcebe46c318cd3",
"submitter": "Karan Ganesan",
"github_login": "karanganesan",
"run_id": "519ac9f9-4bbb-4816-86ea-ac99387c5e1e",
"tier": "easy",
"dataset_id": "e5",
"status": ... | {
"score": {
"mean_loss": 6.209378242492676,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.00375000003259629
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 7.380488395690918,
"example_count": 600,
... |
1af35559-c149-4bdd-9e79-065b715bd19f | easy | oupadhyay | 2026-08-12 07:19:42.485312+00:00 | succeeded | c51991fc6eb6aae05e89099ca4240b090e0e98f43a35185efd1441d74446de1e | 8,373 | null | """Standalone RMS/axial crossing C21 for the Easy E5 tier."""
from __future__ import annotations
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import ModelSpec, OptimizerBundle, OptimizerSpec, Submission, assert_model_state
PAD, N_MARK, X_MARK, T_MARK, DIGIT_BASE =... | {
"id": "1af35559-c149-4bdd-9e79-065b715bd19f",
"created_at": "2026-08-12 07:19:42.485312+00:00",
"db_md5": "ad934b5abe9051b4ed7075984a0d1179",
"submitter": "Ojasw Upadhyay",
"github_login": "oupadhyay",
"run_id": "53233bcc-6b33-473e-9073-9571eb2b434c",
"tier": "easy",
"dataset_id": "e5",
"status": "s... | {
"score": {
"mean_loss": 2.234617917062108,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0066666666915019355
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.1739942236331546,
"example_count": 60... |
1af41ad7-ec6d-4ccd-9b77-d789d735a08e | easy | erdavis0 | 2026-08-26 04:44:51.490039+00:00 | succeeded | 6fe6fdf46bcdc9c76ffef64fbdc6b302f2262f8879ac4d7daea4e62dccf8b390 | 30,354 | null | """Latent-role continuous REV C5: hybrid token CE and sequence mixture.
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... | {
"id": "1af41ad7-ec6d-4ccd-9b77-d789d735a08e",
"created_at": "2026-08-26 04:44:51.490039+00:00",
"db_md5": "14ff631f83062f36483f24b1bd94c4c9",
"submitter": "Ethan",
"github_login": "erdavis0",
"run_id": "aa14eab0-d9d3-48e2-974d-df61e7a42fc6",
"tier": "easy",
"dataset_id": "e10",
"status": "succeeded"... | {
"score": {
"mean_loss": 3.0969956517219543,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.20315152406692505
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 1.2876125574111938,
"example_count": 125... |
1b012c15-3d42-48d4-b654-3e2ce790417c | easy | velocizapkar | 2026-08-16 06:57:17.898336+00:00 | succeeded | b9e5675d8bad174e8328aca21eab824edbcd41d4b04d54d7f4259594da11f877 | 9,535 | null | """Matched standard bidirectional Transformer control.
This is deliberately an ordinary fixed-depth Transformer: learned token and
absolute-position embeddings, full softmax self-attention, pre-RMSNorm,
SwiGLU, residual connections, a final RMSNorm, and an untied vocabulary head.
It has no recurrence, convolution, scr... | {
"id": "1b012c15-3d42-48d4-b654-3e2ce790417c",
"created_at": "2026-08-16 06:57:17.898336+00:00",
"db_md5": "6f8b153263e8d9841b943c21bb6d18f7",
"submitter": "Aakanksh Zarapkar",
"github_login": "velocizapkar",
"run_id": "751978fa-5c74-46fc-9228-81b510c468d7",
"tier": "easy",
"dataset_id": "e6",
"statu... | {
"score": {
"mean_loss": 4.328207969665527,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.02837838069535792
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 4.034871578216553,
"example_count": 60,
... |
1b026c24-a6b1-4915-8e4f-82fde8ae555c | easy | oupadhyay | 2026-08-23 23:21:06.572294+00:00 | succeeded | 61ccd623fdb59a54ba2525876e8cb6194767ca0394b9a508763df344ac589b68 | 7,795 | null | """BiScan entropy preference with twofold T1 endpoint weighting."""
import torch
import torch.nn.functional as F
from torch import nn
from benchmark import OptimizerBundle, Submission, TokenLossBatch, assert_model_state
PAD, N, X, T, ANS, DIGIT = 0, 2, 3, 4, 5, 7
MAX_STEPS = 64
class Config:
def __init__(self,... | {
"id": "1b026c24-a6b1-4915-8e4f-82fde8ae555c",
"created_at": "2026-08-23 23:21:06.572294+00:00",
"db_md5": "ebb1d67196a18440e52cc05b5cd756ff",
"submitter": "Ojasw Upadhyay",
"github_login": "oupadhyay",
"run_id": "60e81b85-3c66-42be-9c14-3b70415cb3e3",
"tier": "easy",
"dataset_id": "e5",
"status": "s... | {
"score": {
"mean_loss": 2.3251134395674313,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.005833333358168602
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.327336972604418,
"example_count": 600... |
1b06d008-89e0-4a49-90d4-f2589a969ed6 | easy | EyimofeA | 2026-08-05 23:16:17.023354+00:00 | succeeded | d7200ea4d2d53af87232083cecf53c4052f246fcdb8ea89928bee7209d9ef51d | 9,071 | null | """Fable Hard arch + train-time T cap + AdamW (timeout-safe).
Merged from `fable_hard_h1_muon` architecture with two changes aimed at the
Hard wall-clock timeout (job 5b363135…, MAX_LOOPS=64 unroll + Muon NS + no
eval_batch_size):
1. **T handling:** parse T from the prompt (same derived features). During
`train()`... | {
"id": "1b06d008-89e0-4a49-90d4-f2589a969ed6",
"created_at": "2026-08-05 23:16:17.023354+00:00",
"db_md5": "df318dd463e79dea4d933da4445d6cad",
"submitter": "mof",
"github_login": "EyimofeA",
"run_id": "258ea325-4ecd-4489-aa99-c807d0696cf4",
"tier": "easy",
"dataset_id": "e1",
"status": "succeeded",
... | {
"score": {
"mean_loss": 1.8563520312309265,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.07999999821186066
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 1.8166258335113525,
"example_count": 100... |
1b08d786-8f5f-4e78-a6f5-cc3043f8e886 | easy | TZZheng | 2026-08-06 21:19:56.352946+00:00 | succeeded | 8ed119744976af18c314e728fd72c5ed9655225dec68493b943a85ac099fa0a6 | 7,300 | 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": "1b08d786-8f5f-4e78-a6f5-cc3043f8e886",
"created_at": "2026-08-06 21:19:56.352946+00:00",
"db_md5": "136b65e97351ed5257a1af84a0ed649c",
"submitter": "TZZheng",
"github_login": "TZZheng",
"run_id": "9114f55c-6be1-4bbb-bba2-caaeba5c085e",
"tier": "easy",
"dataset_id": "e1",
"status": "succeeded"... | {
"score": {
"mean_loss": 2.0441035628318787,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.07999999821186066
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.172515392303467,
"example_count": 100,... |
1b134e67-6a96-4e51-ad17-ebca5e164107 | easy | viridale | 2026-08-25 23:05:44.830303+00:00 | succeeded | b36c583d58bc25f25d3de8e03e26c4e31cff430c85cbd2985b47de7a4276e3c9 | 28,513 | null | """Constraint-posterior bilinear Transformer program.
hypothesis: v913 is three-seed T64/T64 on the single-position N law, but on
the frozen two-position law its training primal is exactly correct from step 475 while
eval is chance because unconditional position argmax deletes a required mixture. Keep
entr... | {
"id": "1b134e67-6a96-4e51-ad17-ebca5e164107",
"created_at": "2026-08-25 23:05:44.830303+00:00",
"db_md5": "b17a290751f84f487fd1dbcb406e33d9",
"submitter": "priormancer",
"github_login": "viridale",
"run_id": "c7c15775-dfec-4bae-93e9-18ddc661f04d",
"tier": "easy",
"dataset_id": "e4",
"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": 1200,
"exact_accuracy": 1.0,
... |
1b151a72-bf21-4522-b729-af06f35031cb | easy | yashkant | 2026-08-19 01:15:00.202339+00:00 | succeeded | 3ae1a37857aee4fed771a0b4c0660f750c47458fcd0c7b8bfc343eb7a30075ed | 45,397 | null | """Round2219 phase-balanced initialization for an anchored recurrent branch.
The confirmed correction-reset parent receives one width-8 recurrent path.
Its live transition and frozen initial anchor execute in parallel, so their
difference is exactly zero at construction and the parent function is kept.
The candidate s... | {
"id": "1b151a72-bf21-4522-b729-af06f35031cb",
"created_at": "2026-08-19 01:15:00.202339+00:00",
"db_md5": "0be9082b1b049610719f6dea1b3e55b4",
"submitter": "Yash Kant",
"github_login": "yashkant",
"run_id": "3f868f3b-03da-4fb2-9144-7b3af15817ba",
"tier": "easy",
"dataset_id": "e1",
"status": "succeed... | {
"score": {
"mean_loss": 4.061444163322449,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.07499999925494194
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 4.297060966491699,
"example_count": 100,
... |
1b222215-9af4-4cf9-ab90-fecc25cff6b1 | easy | sankalp1999 | 2026-08-04 11:32:06.254589+00:00 | succeeded | 05037c83192a2e5e6ac6c9ed75db8771feb7b81239e8fef6f414e21229202909 | 4,988 | null | """E2 control: same flattened-grid MLP, thinner + more steps, betting on grokking.
v1 reached 100% train / 4% test in 7,249 steps. Modular-arithmetic grokking
typically needs far more optimizer steps under strong decoupled weight decay,
so this variant buys steps by shrinking the model (width 768->384, depth 3->2)
and... | {
"id": "1b222215-9af4-4cf9-ab90-fecc25cff6b1",
"created_at": "2026-08-04 11:32:06.254589+00:00",
"db_md5": "2dd3814352cd7f537c2b62260ebe76af",
"submitter": "Sankalp ",
"github_login": "sankalp1999",
"run_id": "ca0a84e1-85b3-4244-9ff9-1ff207716f9d",
"tier": "easy",
"dataset_id": "e1",
"status": "succe... | {
"score": {
"mean_loss": 7.7289862632751465,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0950000025331974
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 8.870741844177246,
"example_count": 100,
... |
1b229b9f-c73f-46a5-a03e-cfe5b9711e22 | easy | erdavis0 | 2026-08-25 01:39:26.891406+00:00 | succeeded | 5adda4144441cbeb2cbe3d83f2248929d1faeef65135607ac4759925de8777bc | 24,108 | null | """Matched categorical-boundary panel: C5 ST oracle robust endpoint.
Public delimiter syntax is used only to route opaque token identities onto
separate right-aligned tapes. One T-blind learned macro maps a categorical X
tape and a digitwise N tape to a new categorical tape through freshly reset
interaction, carry, a... | {
"id": "1b229b9f-c73f-46a5-a03e-cfe5b9711e22",
"created_at": "2026-08-25 01:39:26.891406+00:00",
"db_md5": "70c2cd7e8a73ef8015c8809d1edb6159",
"submitter": "Ethan",
"github_login": "erdavis0",
"run_id": "e60478f2-ccae-4512-b600-637252ee03e5",
"tier": "easy",
"dataset_id": "e3",
"status": "succeeded",... | {
"score": {
"mean_loss": 2.12488306423829,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.010624999999999999
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.1234172756650844,
"example_count": 800,... |
1b2e053d-1a11-41f8-a812-7eb49557ccbd | easy | nikhilkumar26 | 2026-08-31 03:09:55.832249+00:00 | succeeded | 3fe5a5cd0e582802df13f6cfbafb4e8a145e989d482e3c5f52d4287ce5d012eb | 18,979 | null | """T-adaptive learned recurrence. T is unroll count only.
A shared step maps digit embeddings of the current integer and of N to
next-digit logits. Between unrolls, STE/argmax updates the digit state so T
composes. No problem-specific operators are hardcoded.
"""
from __future__ import annotations
import math
impor... | {
"id": "1b2e053d-1a11-41f8-a812-7eb49557ccbd",
"created_at": "2026-08-31 03:09:55.832249+00:00",
"db_md5": "d3b25d4107294cd212bac47d953756a2",
"submitter": "Nikhil Kumar",
"github_login": "nikhilkumar26",
"run_id": "c69eec5a-95c6-4c7a-9544-0137e35d3c4e",
"tier": "easy",
"dataset_id": "e1",
"status": ... | {
"score": {
"mean_loss": 2.6543865890174945,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.05666666680326064
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.7180788837262053,
"example_count": 100... |
1b3051e3-ead7-41b8-b7b5-5522f3479832 | easy | gauravmishra | 2026-08-08 02:03:21.547531+00:00 | succeeded | adcc65d01978fe5961c3181b18a0e7c523092969f7a0452f93553e5543ffd75b | 3,881 | null | """Four-pass Transformer with two blocks applied in an A-B-A-B pattern."""
from __future__ import annotations
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
assert_model_state,
)
D_MODEL = ... | {
"id": "1b3051e3-ead7-41b8-b7b5-5522f3479832",
"created_at": "2026-08-08 02:03:21.547531+00:00",
"db_md5": "3340cb471ec609c7e01b51846d4354b3",
"submitter": "Gaurav Mishra",
"github_login": "gauravmishra",
"run_id": "46a09a37-8959-4a8c-97f1-6718f312eb7c",
"tier": "easy",
"dataset_id": "e5",
"status": ... | {
"score": {
"mean_loss": 4.9443469120125805,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.009583333345750969
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 5.304946984945391,
"example_count": 600... |
1b31d718-1d24-40ec-99ba-2593ff278a39 | easy | newjordan | 2026-08-22 19:13:22.688174+00:00 | succeeded | d3de64b30978fd5a517d936bb801964b6f4cce6e207a77d82c3cc7eb410ee4b2 | 24,338 | null | """Four-lane register Transformer with a tied learned transition.
Decimal fields are packed into a fixed-width little-endian register. State,
modulus, work, and scratch lanes are refined by one shared block. A learned
ordered-pair relation tensor gives WORK/SCRATCH one cross-attention
initialization from the current... | {
"id": "1b31d718-1d24-40ec-99ba-2593ff278a39",
"created_at": "2026-08-22 19:13:22.688174+00:00",
"db_md5": "1f6769056c31fc795a9a8815d6029929",
"submitter": "Frosty40",
"github_login": "newjordan",
"run_id": "58912a94-c9d8-4b8a-bde1-aec2ae7fe6fe",
"tier": "easy",
"dataset_id": "e6",
"status": "succeed... | {
"score": {
"mean_loss": 2.1287935972213745,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.25788289308547974
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.0201902389526367,
"example_count": 60,... |
1b31fd22-2a74-4548-8b79-52d1fadb17a6 | easy | amirmeisami | 2026-08-15 21:37:08.691876+00:00 | succeeded | 9712fa2cb35b9e5ba3aab999f10071f17a0c2e5076d1ecb58563fb29cfe6fa26 | 15,479 | null | """E010-A0: state-causal shared-transition recurrent 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,
assert_mod... | {
"id": "1b31fd22-2a74-4548-8b79-52d1fadb17a6",
"created_at": "2026-08-15 21:37:08.691876+00:00",
"db_md5": "f1a4ab71de668a0eb58e1aa6fd48a63d",
"submitter": "Amir Meisami",
"github_login": "amirmeisami",
"run_id": "a2b752ab-8120-4b2f-91f5-1e121864fd3c",
"tier": "easy",
"dataset_id": "e6",
"status": "s... | {
"score": {
"mean_loss": 5.7589214827916395,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.08198198595562498
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 5.808387279510498,
"example_count": 60,
... |
1b368e87-b3ef-4254-a653-ebb8cefa0cbc | easy | DDanlov | 2026-08-13 19:46:34.232855+00:00 | succeeded | 939f58d818c3f953053dfcbc068030b31142e50f323a87b6d7f84ab5df0f2dd5 | 26,904 | null | from __future__ import annotations
import math
import torch
import torch.nn as nn
import torch.nn.functional as F
from torch.optim.optimizer import Optimizer
try:
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
assert_model_state,
)
except ... | {
"id": "1b368e87-b3ef-4254-a653-ebb8cefa0cbc",
"created_at": "2026-08-13 19:46:34.232855+00:00",
"db_md5": "e3de2d9e8ccead933e2b8372c4ddba6b",
"submitter": "DDanlov",
"github_login": "DDanlov",
"run_id": "3bd6b3ec-88fd-4d57-8fb7-c7204e76c992",
"tier": "easy",
"dataset_id": "e3",
"status": "succeeded"... | {
"score": {
"mean_loss": 2.3031981193112196,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.018125000000000002
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.303733025772711,
"example_count": 800... |
1b394535-9840-47e2-a64c-6958045d086e | easy | yunjiangster | 2026-08-28 05:11:29.309299+00:00 | succeeded | 287be7747a2fe069350f84c5781e1f499ed35f2cb4f4c69e87241c77c6d3ee07 | 15,492 | null | """Continuous shared neural transition experiment for One Layer Deeper.
This model intentionally does *not* contain a symbolic arithmetic program or a
hard-coded squaring/modulo rule. It uses the documented Easy/Medium prompt
format only as an interface to obtain N, x, and T, then learns a single neural
transition F_... | {
"id": "1b394535-9840-47e2-a64c-6958045d086e",
"created_at": "2026-08-28 05:11:29.309299+00:00",
"db_md5": "0c323eca4c94d16e1ca36fa2b67082f5",
"submitter": "Yunjiang Jiang",
"github_login": "yunjiangster",
"run_id": "1302a5cd-800f-4a8a-b416-38408be4ba1f",
"tier": "easy",
"dataset_id": "e1",
"status":... | {
"score": {
"mean_loss": 2.0521366000175476,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.08999999985098839
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 1.8170844316482544,
"example_count": 100... |
1b3c53f5-d96d-4e4b-a752-6bb2342644f2 | easy | erdavis0 | 2026-08-29 00:49:59.575191+00:00 | succeeded | 6dd63d7fbd57f1dfcb99a68f4eb1c3004d5da2334182db359d857967577b9df9 | 20,131 | null | """Curvature table with a compact gentle modulus-relative residual."""
from __future__ import annotations
import math
import time
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
TokenLossBatc... | {
"id": "1b3c53f5-d96d-4e4b-a752-6bb2342644f2",
"created_at": "2026-08-29 00:49:59.575191+00:00",
"db_md5": "02d94eaf51f2ef55a1f2bce028a4c985",
"submitter": "Ethan",
"github_login": "erdavis0",
"run_id": "c020e58e-59aa-46f9-b49d-7b9b90461c1b",
"tier": "easy",
"dataset_id": "e2",
"status": "succeeded",... | {
"score": {
"mean_loss": 0.10724670712693296,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.987500011920929
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 4.773826276505133e-06,
"example_count": 3... |
1b420eb4-27e7-44ff-84ce-4c478a7b82e5 | easy | amirmeisami | 2026-08-16 00:18:26.906813+00:00 | succeeded | 2422a5f6cbbae2d5be212088e9b82c510a2bef1cd1058a5faa830bbd35c93d83 | 21,274 | null | """E016: context-attentive closed-state digit transition."""
from __future__ import annotations
import math
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
TokenLossBatch,
assert_model_st... | {
"id": "1b420eb4-27e7-44ff-84ce-4c478a7b82e5",
"created_at": "2026-08-16 00:18:26.906813+00:00",
"db_md5": "b24fcfa2715bad3c3f013dce36d9f065",
"submitter": "Amir Meisami",
"github_login": "amirmeisami",
"run_id": "d821f289-5464-4eac-8031-6b1f5e05dfa3",
"tier": "easy",
"dataset_id": "e5",
"status": "s... | {
"score": {
"mean_loss": 2.3319886425849496,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0033333333333333335
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.3862699842239175,
"example_count": 6... |
1b429ee3-bed1-4175-ad9a-c6fa56562405 | easy | isaac0804 | 2026-08-05 08:30:54.250406+00:00 | succeeded | 0ab40eb7cfa42b9700f7e4a4f170eba07ae49a93a935f124b68d54c84d706635 | 9,365 | null | """Weight-tied looped Transformer (v6 core: damped gate, 2 sublayers/iteration)
plus explicit place-value embeddings for N's and x's digit tokens.
Motivation (see docs/findings.md, N-conditioning section): every architecture
tried so far treats N's (and x's) digits as an opaque span for generic
attention to notice imp... | {
"id": "1b429ee3-bed1-4175-ad9a-c6fa56562405",
"created_at": "2026-08-05 08:30:54.250406+00:00",
"db_md5": "6c332bc632593f83d44af5125cddee51",
"submitter": "Isaac Yong",
"github_login": "isaac0804",
"run_id": "d4943e16-2359-41b8-ae9d-dbf1fe946542",
"tier": "easy",
"dataset_id": "e5",
"status": "succe... | {
"score": {
"mean_loss": 2.203705829171154,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.009166666679084301
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.210826189528666,
"example_count": 600,... |
1b5224dd-4fe3-4d43-8e5c-0ed807cfc751 | easy | hamzah2304 | 2026-08-24 18:57:46.930303+00:00 | succeeded | dd26eba4da94876bb09601940e3a6f124ee2694e6f272108cce02044af9aefef | 5,424 | 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
PA... | {
"id": "1b5224dd-4fe3-4d43-8e5c-0ed807cfc751",
"created_at": "2026-08-24 18:57:46.930303+00:00",
"db_md5": "e70ebc69a4c598b060cdd090a8393218",
"submitter": "hamzah2304",
"github_login": "hamzah2304",
"run_id": "7e1eb733-9abb-4901-83d6-b570de436bf8",
"tier": "easy",
"dataset_id": "e1",
"status": "succ... | {
"score": {
"mean_loss": 2.0735466480255127,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.05166666489094496
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.0696771144866943,
"example_count": 100... |
1b5db94a-7472-4ea7-9334-7cf5e9eebad6 | easy | Ranoobaba | 2026-08-16 06:06:13.568468+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": "1b5db94a-7472-4ea7-9334-7cf5e9eebad6",
"created_at": "2026-08-16 06:06:13.568468+00:00",
"db_md5": "eaac69fe345f82973fcf392c504a40ff",
"submitter": "Syed Rayyan Ali",
"github_login": "Ranoobaba",
"run_id": "e45b0244-0ea8-4bec-a483-780c17db4e6f",
"tier": "easy",
"dataset_id": "e1",
"status": "... | {
"score": {
"mean_loss": 2.9941506385803223,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.008333333302289248
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 3.9741852283477783,
"example_count": 10... |
1b613f49-68b7-4895-a7e6-6f0f0ccabf76 | easy | chuk-uzowihe | 2026-08-08 02:47:17.785583+00:00 | succeeded | 976eb35378786317b47984fcc59bf216633b68f5fea8decd194f579c1d853759 | 27,744 | null | """One Layer Deeper submission: FPRM over a compressed (N, x, T) lattice.
v6 = v2's architecture (the only version that has generalized on m1: 15.18%
test, 11.6% at unseen T=32) rebuilt on the v5 chassis, with its one fatal
flaw fixed: input numbers are no longer parsed into int64 (silent garbage
past 19 digits) — ins... | {
"id": "1b613f49-68b7-4895-a7e6-6f0f0ccabf76",
"created_at": "2026-08-08 02:47:17.785583+00:00",
"db_md5": "ea3c3d0462cf17d2298bb1384f116309",
"submitter": "Chuk Uzowihe",
"github_login": "chuk-uzowihe",
"run_id": "f9251c19-8f2d-4c32-86ab-a9ce4a1960a1",
"tier": "easy",
"dataset_id": "e4",
"status": "... | {
"score": {
"mean_loss": 14.0805578112899,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.004166666666666667
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 14.512738526966302,
"example_count": 1200... |
1b61c7dd-4465-4889-b8f8-767a7b0c12cd | easy | sapient-sapiens | 2026-08-14 18:38:03.949964+00:00 | succeeded | 28ba469aae3e42ca679270105f219b6bd47433fa80f83dc154b3c3e55e44c7c3 | 14,299 | null | """Digit-space recurrent operator for repeated modular squaring.
Nothing here converts a token into a decimal number. Digits stay as embeddings,
so what a digit is worth, and every carry and reduction, is learned. That also
sidesteps a hard numeric wall: at Hard's modulus scale `x^2` runs past the 53-bit
exact-integer... | {
"id": "1b61c7dd-4465-4889-b8f8-767a7b0c12cd",
"created_at": "2026-08-14 18:38:03.949964+00:00",
"db_md5": "1efadda9e01ae84ec2857432a666c542",
"submitter": "Ertondy",
"github_login": "sapient-sapiens",
"run_id": "1c4d63e2-fd84-49ad-96a1-0ac80706264a",
"tier": "easy",
"dataset_id": "e3",
"status": "su... | {
"score": {
"mean_loss": 2.1322567870604825,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.011250000000000001
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.129992060867939,
"example_count": 800... |
1b64782e-a072-483f-ae29-1d9da924ad54 | easy | MarcoF1 | 2026-08-13 16:49:49.407687+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": "1b64782e-a072-483f-ae29-1d9da924ad54",
"created_at": "2026-08-13 16:49:49.407687+00:00",
"db_md5": "498f495d8a0b910efb7ca633ca25eee2",
"submitter": "Marco Fleming",
"github_login": "MarcoF1",
"run_id": "72961337-116a-4589-8b7b-8f7c0041e9cb",
"tier": "easy",
"dataset_id": "e1",
"status": "succ... | {
"score": {
"mean_loss": 2.770329301196738,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.07500000024835268
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 3.200248956680298,
"example_count": 100,
... |
1b682d0c-ca7b-4ac7-addc-ce8c7615959f | easy | gauravmishra | 2026-08-09 19:09:01.671050+00:00 | succeeded | 58416b2e4c40aa8ba8fb5b75344a1fd30868f078df98b2e9fd6fe9d870d05e9b | 4,268 | 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,
TokenLossBatch,
assert_model_state,
)
D_MODEL = ... | {
"id": "1b682d0c-ca7b-4ac7-addc-ce8c7615959f",
"created_at": "2026-08-09 19:09:01.671050+00:00",
"db_md5": "2f65708c108a34a49bbdfae7be9de9e6",
"submitter": "Gaurav Mishra",
"github_login": "gauravmishra",
"run_id": "c85010c9-6453-426c-8b1c-5e5db015f0fc",
"tier": "easy",
"dataset_id": "e5",
"status": ... | {
"score": {
"mean_loss": 2.2721646196545144,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.007500000012417634
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.2883040006919826,
"example_count": 60... |
1b6c15fc-8af4-48e5-b968-7952d08af98c | easy | ehonig | 2026-08-22 04:40:06.693892+00:00 | succeeded | 44ecd8da499c1559614d613737c668bfd0b20a76f20906717baa9072f4b6ae8a | 31,880 | 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": "1b6c15fc-8af4-48e5-b968-7952d08af98c",
"created_at": "2026-08-22 04:40:06.693892+00:00",
"db_md5": "505e99fe53a2fa4bed673971201f22cc",
"submitter": "Edouardo Honig",
"github_login": "ehonig",
"run_id": "2f66e177-c52e-425b-8d0e-8042f4f8852e",
"tier": "easy",
"dataset_id": "e1",
"status": "succ... | {
"score": {
"mean_loss": 2.2281529517047667,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.08333333387970925
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.2405840078396584,
"example_count": 100... |
1b78c440-6993-467b-bf9e-b38a93a9782b | easy | alirezashirvani-jr | 2026-08-28 23:33:27.893374+00:00 | succeeded | 73dad23ac67ce34a8f5ba4257342fbd955f792c46dd910f2668b969b51df72ec | 53,560 | 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": "1b78c440-6993-467b-bf9e-b38a93a9782b",
"created_at": "2026-08-28 23:33:27.893374+00:00",
"db_md5": "c68d20cdc8ff54eb88edee065db13428",
"submitter": "alirezashirvani-jr",
"github_login": "alirezashirvani-jr",
"run_id": "61bf2f70-7f6a-464f-9d23-8ad76c186f79",
"tier": "easy",
"dataset_id": "e4",
... | {
"score": {
"mean_loss": 2.109929317739425,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.008020833345750968
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.113878182766149,
"example_count": 1200... |
1b7da573-1086-4da1-942b-7052fa09bf29 | easy | jordanrubin | 2026-08-12 07:11:37.420454+00:00 | succeeded | 60a2a54debdbc8b340d199e9bf0aa3ebed2491bdc562ee97b5c6f1b19eda4130 | 24,919 | 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": "1b7da573-1086-4da1-942b-7052fa09bf29",
"created_at": "2026-08-12 07:11:37.420454+00:00",
"db_md5": "c1ea7b63327dc7beb4d7cf3e3c9676a0",
"submitter": "Jordan Rubin",
"github_login": "jordanrubin",
"run_id": "e6c9d8dc-665c-47b9-a090-531b4277e7b1",
"tier": "easy",
"dataset_id": "e5",
"status": "s... | {
"score": {
"mean_loss": 2.9382113618344645,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.02916666547146936
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.2449243372064,
"example_count": 600,
... |
1b7ed9d6-7a5a-46ad-9c1c-7088bf48732c | easy | ddoan | 2026-08-23 19:56:07.002039+00:00 | failed | 8885cf0b067b55b0803ccba1b00c9a19f1abef090dbabe4b3a1286731eee0ed9 | 45,976 | null | """Parallel semantic learners with an exact 62-bit execution tournament.
Each expert learns independently from the supplied endpoint loss. A second
learned categorical field tracks which expert currently explains the training
rows best. Evaluation executes the selected expert and its selected program
for as many as ... | {
"id": "1b7ed9d6-7a5a-46ad-9c1c-7088bf48732c",
"created_at": "2026-08-23 19:56:07.002039+00:00",
"db_md5": "b35b7c8228f01580a18aa09a9a742efe",
"submitter": "Doug Doan",
"github_login": "ddoan",
"run_id": "d1b8e061-94f9-474d-a317-b4dec4d0f75f",
"tier": "easy",
"dataset_id": "e1",
"status": "failed",
... | null
|
1b850694-1da7-44e9-95de-73bb6fa6df3b | easy | khushidahi | 2026-08-22 01:31:14.988978+00:00 | failed | 77a5769dec2ffebc44376909623880d20423acf8f5c2c3170e631656bcbaaa8a | 19,214 | null | """R395: minimal Universal-Transformer functional-root control for M5.
This candidate adapts the released Universal Transformer to the competition's
separate-output interface. Numeric prompt symbols are only rearranged onto
equal-length, right-aligned categorical N and X tapes. A third tape supplies
parallel learned... | {
"id": "1b850694-1da7-44e9-95de-73bb6fa6df3b",
"created_at": "2026-08-22 01:31:14.988978+00:00",
"db_md5": "ffa10c073262fa9b74880c658acc63ec",
"submitter": "khushidahi",
"github_login": "khushidahi",
"run_id": "57c0e887-0cc3-40fd-b361-1b371d2ab09e",
"tier": "easy",
"dataset_id": "e5",
"status": "fail... | null
|
1b883d94-30ad-4601-a2c4-b2b01e7e02f8 | easy | SunstarKK | 2026-08-18 12:38:09.089008+00:00 | succeeded | 85ad787404d6972c3a77e3033ac814214af05355e61d895e6d3f66a586b2909c | 13,775 | null | """Learned multiscale digit pyramid for One Layer Deeper."""
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,
)
PAD ... | {
"id": "1b883d94-30ad-4601-a2c4-b2b01e7e02f8",
"created_at": "2026-08-18 12:38:09.089008+00:00",
"db_md5": "e026ee63e238218eaba8ff87632f53d3",
"submitter": "SunstarK",
"github_login": "SunstarKK",
"run_id": "df8da52a-a53c-4bae-a431-c0e5761b08ea",
"tier": "easy",
"dataset_id": "e5",
"status": "succeed... | {
"score": {
"mean_loss": 2.7833427572593923,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.010416666716337205
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.3918699523258637,
"example_count": 60... |
1b89bee5-793b-4d6f-8d52-842cb7f9a1a1 | easy | sapient-sapiens | 2026-08-17 18:50:09.506872+00:00 | succeeded | 36f388e7773a10c6e9f21218c3b2b7c2e02ef662b80609c9e89d85f1a829c2bf | 28,749 | 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": "1b89bee5-793b-4d6f-8d52-842cb7f9a1a1",
"created_at": "2026-08-17 18:50:09.506872+00:00",
"db_md5": "3a4be9949d77b29e94511b005604756c",
"submitter": "Ertondy",
"github_login": "sapient-sapiens",
"run_id": "e78b7caa-1aeb-4fe4-8118-d4ddba21ec67",
"tier": "easy",
"dataset_id": "e9",
"status": "su... | {
"score": {
"mean_loss": 1.8976770378180055,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.02666666753590107
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 1.8984602689743042,
"example_count": 90,... |
1b95af05-6f91-42b0-9ed7-f716272a0ffc | easy | yashkant | 2026-08-27 23:36:48.386486+00:00 | succeeded | 73442cd87ac8dbc4a411706e3a74bc07a4549b186c81b10ee43b9b06851f6ae9 | 29,617 | null | """v16rr = identical copy of v16r (repeatability run).
v16r = v16 (mixed-scale operands) + hard-example replay for the squarer.
One Layer Deeper HARD submission v16 -- EVERY ARITHMETIC DECISION LEARNED: a learned squarer (parameter-free
column routing + a small transformer (default) or one tied carry cell, trained on... | {
"id": "1b95af05-6f91-42b0-9ed7-f716272a0ffc",
"created_at": "2026-08-27 23:36:48.386486+00:00",
"db_md5": "4dd63ddbfa4c221ad5d1b83aff0d09f3",
"submitter": "Yash Kant",
"github_login": "yashkant",
"run_id": "299181ca-ea23-48ee-a508-aa1081fb2aee",
"tier": "easy",
"dataset_id": "e5",
"status": "succeed... | {
"score": {
"mean_loss": 0.14880771834512352,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.9920833388964335
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 0.17937219142913818,
"example_count": 60... |
1b9a4da8-d64d-4825-a46a-3394bd04ec03 | easy | rougee | 2026-08-26 03:04:09.182355+00:00 | succeeded | 10e18ab2b9aec2aa59ad47c50c9697b360073fa301f5131b2f64a824aabcbb3f | 12,437 | null | """Official Submittable Champion for One Layer Deeper (Polynomial Quadratic Recurrence + Cross-Attention Readout).
State-of-the-Art Recurrent Architecture:
- Prelude: Bidirectional Multi-Head Self-Attention Contextual Prompt Encoder with Forward & Reverse Positional Embeddings
- Metric Geometry: Dynamic Skew-Symmetric... | {
"id": "1b9a4da8-d64d-4825-a46a-3394bd04ec03",
"created_at": "2026-08-26 03:04:09.182355+00:00",
"db_md5": "5e06ce26b1cc03b8f43fc20b270850a5",
"submitter": "Rougee",
"github_login": "rougee",
"run_id": "d4405d9f-ea9f-431a-a4df-0133221509d0",
"tier": "easy",
"dataset_id": "e1",
"status": "succeeded",
... | {
"score": {
"mean_loss": 10.327543816364317,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.044999998509883885
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 11.364618301391602,
"example_count": 10... |
1b9bd3cb-ad3d-4040-b02b-e35de73af91e | easy | gauravmishra | 2026-08-08 00:04:33.459547+00:00 | succeeded | 9140fef522ec6a9afeb558e69343f534c498c0164ab7648853f9bbf90a742043 | 3,726 | null | """Four-pass Transformer with four independently parameterized blocks."""
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 = 1... | {
"id": "1b9bd3cb-ad3d-4040-b02b-e35de73af91e",
"created_at": "2026-08-08 00:04:33.459547+00:00",
"db_md5": "067dd19f95b3a2b025d36561c727f6b0",
"submitter": "Gaurav Mishra",
"github_login": "gauravmishra",
"run_id": "4dedc7b3-d89b-4b6b-87fe-be58252b2e4b",
"tier": "easy",
"dataset_id": "e5",
"status": ... | {
"score": {
"mean_loss": 11.379916662725565,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.006666666716337203
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 11.57419835505464,
"example_count": 600... |
1ba28b21-8164-463a-abaa-335ef6fb1e5d | easy | jordanrubin | 2026-08-05 21:02:56.992405+00:00 | succeeded | d16f7833417394a174ed8b04a661869b08d445d0b92a047a2dcc58f88ec2c3ed | 24,859 | null | """Adaptive exploration template for the exact-T looped Transformer.
The T=1 computation is the proven value-Fourier baseline cell: a D=256,
three-block bidirectional Transformer with abacus digit positions and learned
Fourier features over the parsed integer x. For T>1 the *same* three-block
cell is reused exactly T... | {
"id": "1ba28b21-8164-463a-abaa-335ef6fb1e5d",
"created_at": "2026-08-05 21:02:56.992405+00:00",
"db_md5": "495361eb0a822d607d725e29b331305f",
"submitter": "Jordan Rubin",
"github_login": "jordanrubin",
"run_id": "5630c70a-6679-48c2-b341-dd2cc3b7bb24",
"tier": "easy",
"dataset_id": "e1",
"status": "s... | {
"score": {
"mean_loss": 5.534323304965906,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.2283333291610082
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 6.915815353393555,
"example_count": 100,
... |
1ba29186-63e3-438b-a7ae-d83c15e2a084 | easy | velocizapkar | 2026-08-04 06:19:41.159526+00:00 | failed | c89d45d0e4e814826d965a716ba2a1ef41599256fb3401079036ba9e61c9fa2f | 42,649 | null | """Orthogonal depth lattice with recurrent Block AttnRes.
The learned state has three explicit axes: sequence/work position, parallel
computational slot, and recurrent reasoning time. The working transition is
a composition of Householder mixers along the first two axes and invertible
radial twists along features. A... | {
"id": "1ba29186-63e3-438b-a7ae-d83c15e2a084",
"created_at": "2026-08-04 06:19:41.159526+00:00",
"db_md5": "b72f8e1adc08719c3fc270e0978764bc",
"submitter": "Aakanksh Zarapkar",
"github_login": "velocizapkar",
"run_id": "fb4ef2b3-cb2a-4241-9694-b0baaa2cb039",
"tier": "easy",
"dataset_id": "e3",
"statu... | null
|
1ba76524-db41-4150-b2fb-b0d4092482d5 | easy | yashkant | 2026-08-17 00:36:10.389069+00:00 | succeeded | fbe498096c6fd10839797bcafeaa9f2ab1a4f9d407244b6c3088236992be104a | 43,303 | null | """Round2213 learns a channel-wise reset decision from fixed context.
The confirmed parent resets both width-8 Dykstra correction memories on
every parsed-T cycle. One shared zero-start 9-to-8 affine gate sees the
immutable pooled input/answer anchor and explicit parsed T each cycle.
Its clipped channel outputs inter... | {
"id": "1ba76524-db41-4150-b2fb-b0d4092482d5",
"created_at": "2026-08-17 00:36:10.389069+00:00",
"db_md5": "d6c00a1aff16a0c4dc3c60364f24bdff",
"submitter": "Yash Kant",
"github_login": "yashkant",
"run_id": "8a0cc09a-1d3a-4aca-b2ed-741ea0e8c3d4",
"tier": "easy",
"dataset_id": "e3",
"status": "succeed... | {
"score": {
"mean_loss": 2.1323036128962993,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.016875000167638064
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.132715551904334,
"example_count": 800... |
1ba96988-a443-4a2c-b2df-89b53a43c9d8 | easy | oupadhyay | 2026-08-25 03:36:14.755738+00:00 | succeeded | 6f72d75024989489d3206dc9e6e175f6af9aa67d176dd2f2a9a40bd6b5e2f673 | 3,921 | null | """Ordered pair mixer: efficient global interactions for Easy depth learning."""
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,R,H=4,16,16,64
class C:
def __init__(s,vocab_size,... | {
"id": "1ba96988-a443-4a2c-b2df-89b53a43c9d8",
"created_at": "2026-08-25 03:36:14.755738+00:00",
"db_md5": "b9e5c1923e22565e7e13e91536c31e9b",
"submitter": "Ojasw Upadhyay",
"github_login": "oupadhyay",
"run_id": "d9cc1ff6-8028-4fc6-9876-cfa425bf48e3",
"tier": "easy",
"dataset_id": "e5",
"status": "s... | {
"score": {
"mean_loss": 2.1613937374194614,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.008750000012417634
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.1410823841266033,
"example_count": 60... |
1bae46c1-13ed-46db-b7cc-cd2bdadf789f | easy | yashkant | 2026-08-19 01:40:05.189220+00:00 | succeeded | 22b48a4860926826a1d5de2844dbe04a937b1b2a5f2174afc64332e2aa050cf4 | 49,530 | null | """Round2217 adds a zero-start square/reduce/answer recurrence.
The confirmed correction-reset model remains intact. Its width-8
answer state is refined by one shared transition with explicit product,
reduction, and answer roles. Immutable categorical N and X summaries
are reinjected on every call and physical calls... | {
"id": "1bae46c1-13ed-46db-b7cc-cd2bdadf789f",
"created_at": "2026-08-19 01:40:05.189220+00:00",
"db_md5": "054a7a506a679f95413f32c3db1665d8",
"submitter": "Yash Kant",
"github_login": "yashkant",
"run_id": "2f1f1f00-5b2e-43f0-b70a-1b414d79560b",
"tier": "easy",
"dataset_id": "e3",
"status": "succeed... | {
"score": {
"mean_loss": 2.1258206614356574,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.018125000093132258
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.127241057912291,
"example_count": 800... |
1baebe09-2764-4dd8-8b56-1ba68bc1cfdb | easy | CodeReclaimers | 2026-08-20 15:17:08.716008+00:00 | succeeded | ecdc5bc3549f21ec4b5703f65802288e92bc05cd2c1587f9346ab4e680aba37d | 42,274 | null | """v39_keyed: keyed-modulus CRT learner — Run A of T1_CERT_PLAN_20260819.
HYPOTHESIS A (alias modulus): h1's displayed N is a stable per-identity
alias phi(M) of a hidden modulus M <= 2^20 from enumerated factor pairs
(<=10-bit primes), while x and y are displayed plainly and the step closes
per prime channel (squarin... | {
"id": "1baebe09-2764-4dd8-8b56-1ba68bc1cfdb",
"created_at": "2026-08-20 15:17:08.716008+00:00",
"db_md5": "4734e439309d100106007d6b654ccab0",
"submitter": "CodeReclaimers",
"github_login": "CodeReclaimers",
"run_id": "73692b13-e223-40f5-b21d-43dc1d8f59aa",
"tier": "easy",
"dataset_id": "e1",
"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": 100,
"exact_accuracy": 1.0,
"... |
1bb3afc5-8e42-491d-afc1-2253daa3c19c | easy | khushidahi | 2026-08-06 04:03:47.122482+00:00 | succeeded | d5526d168dec8697171d9f28b2aaa37d8185ca5d550565934f93cf5b1826544d | 14,402 | 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": "1bb3afc5-8e42-491d-afc1-2253daa3c19c",
"created_at": "2026-08-06 04:03:47.122482+00:00",
"db_md5": "6c534b640bc36f3434ed6ed881f61679",
"submitter": "khushidahi",
"github_login": "khushidahi",
"run_id": "36b2e5e1-5e87-4c59-bee1-f0e0392398ee",
"tier": "easy",
"dataset_id": "e5",
"status": "succ... | {
"score": {
"mean_loss": 2.1727752685546875,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.010416666977107525
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.170254707336426,
"example_count": 600... |
1bb48c8e-f90e-45bc-b05f-33947f0d20ee | easy | DDanlov | 2026-08-15 14:30:52.460491+00:00 | succeeded | cbcce9af75eb899a1ec21cdc9ddf417815202dabb260d8049bdb09de4a3d8dad | 18,987 | null | """
Official Submission for One Layer Deeper Challenge
Architecture: Pure No-Norm, No-Skip DKS Recurrent Core with Double Embedding (dim=512),
Input RMS Normalization only, 64-Channel Harmonic+Dyadic Torus, Power-of-10 Alignment, and RohanShampoo.
"""
from __future__ import annotations
import math
from typing import O... | {
"id": "1bb48c8e-f90e-45bc-b05f-33947f0d20ee",
"created_at": "2026-08-15 14:30:52.460491+00:00",
"db_md5": "3cf72178d359ca91154237c81fe73610",
"submitter": "DDanlov",
"github_login": "DDanlov",
"run_id": "d4f62acf-0c48-4caf-81d4-c2a7288d676c",
"tier": "easy",
"dataset_id": "e5",
"status": "succeeded"... | {
"score": {
"mean_loss": 2.2856328793368794,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.004166666666666667
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.2698144065425,
"example_count": 600,
... |
1bb82512-d652-401c-9408-477c26fba6ff | easy | Yalyenea | 2026-08-12 07:41:17.480783+00:00 | succeeded | 209a2c4ff470ecaee9c76f3ec2b319bd0b9c4c1799b76e2125fe9e0b6684d4aa | 12,139 | null | """Compact decimal recurrence with Muon-optimized transition matrices."""
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,
as... | {
"id": "1bb82512-d652-401c-9408-477c26fba6ff",
"created_at": "2026-08-12 07:41:17.480783+00:00",
"db_md5": "bb642a13202d17077398aaa17ef5ca9b",
"submitter": "yfff",
"github_login": "Yalyenea",
"run_id": "09d10882-ff05-4ced-b7fa-150ecc916737",
"tier": "easy",
"dataset_id": "e5",
"status": "succeeded",
... | {
"score": {
"mean_loss": 5.866601284845128,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.009583333333333333
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 6.354399435188738,
"example_count": 600,... |
1bc75636-48f4-47f5-9604-430902ac3d59 | easy | yashkant | 2026-08-12 03:31:23.312986+00:00 | succeeded | bc49285cdb49d5f24f266aa1cf2ca8c0735143f0fa8ac2bd38b8053a708bff3d | 31,388 | null | """Round1702 one-projection-per-T Dykstra recursion on the D416 carrier.
Each parsed public-T cycle applies one shared learned smooth projection and
alternates the A/B set code and correction memory across cycles. The immutable
anchor is reinjected on every call; all active cycle outputs are averaged, and
terminal CE... | {
"id": "1bc75636-48f4-47f5-9604-430902ac3d59",
"created_at": "2026-08-12 03:31:23.312986+00:00",
"db_md5": "72e2c85438f1d7eb980c6e8e635c03ab",
"submitter": "Yash Kant",
"github_login": "yashkant",
"run_id": "e3662c3f-68e9-4407-b466-63f8144e1c40",
"tier": "easy",
"dataset_id": "e2",
"status": "succeed... | {
"score": {
"mean_loss": 2.564311385154724,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.07979167252779007
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 3.2558746337890625,
"example_count": 300,... |
1bc9f143-64d0-449b-9336-b4ea6c06c280 | easy | AFeinfield | 2026-08-27 22:14:46.248900+00:00 | succeeded | d413b8a9a7b741fca2b70f94caa8ec024e6a344955525a767dceeb080196033d | 12,944 | null | """154: 152 + INTEGERNESS PRIOR on intermediate digit lanes
(0.1 * sin^2(pi*y), zero at integers — generic regularizer).
151 notes: throughput/learnability pass.
Dynamic alignments (m <= batch N-width), carry passes 6 post-square / 3
in reduction, single correction try, explicit y/10 feature (carry is
linear in it), ... | {
"id": "1bc9f143-64d0-449b-9336-b4ea6c06c280",
"created_at": "2026-08-27 22:14:46.248900+00:00",
"db_md5": "385732bbcee47da6be26329c9ddb962f",
"submitter": "AFeinfield",
"github_login": "AFeinfield",
"run_id": "50e2cc8b-9dc1-4942-bb48-b9cedcc62e56",
"tier": "easy",
"dataset_id": "e5",
"status": "succ... | {
"score": {
"mean_loss": 2.239359618720955,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0008333333457509677
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.239800791034784,
"example_count": 600... |
1bd49e14-b344-4ec2-90a4-10e901dd6291 | easy | G-AshwinKumar | 2026-08-14 08:40:31.780984+00:00 | succeeded | e6c60969c4531b9be1df417b408c1ca8847ad4631f2c8d4664148850f4d94f3a | 33,403 | null | """Recurrent squaring operator with a discrete residue state between iterations.
Continuous recurrent states drift: small errors in the learned dynamics compound
over iterations, so a model correct at its training depth degrades when run
longer. The measured fix is to make the state between iterations *discrete* --
re... | {
"id": "1bd49e14-b344-4ec2-90a4-10e901dd6291",
"created_at": "2026-08-14 08:40:31.780984+00:00",
"db_md5": "72669a244c481a0a8cb7297096e19486",
"submitter": "Ashwin Kumar",
"github_login": "G-AshwinKumar",
"run_id": "7950c7b4-06a3-4f00-9d76-a78c31e7f15d",
"tier": "easy",
"dataset_id": "e1",
"status": ... | {
"score": {
"mean_loss": 2.3041363954544067,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.11499999836087227
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.225179672241211,
"example_count": 100,... |
1bdecbdf-39f6-437a-9747-af2fb0cc9ce2 | easy | chad-atexpedient | 2026-08-14 02:49:00.780924+00:00 | succeeded | fd5171acc931f01603fea3bba0e577d3e026ac5a1b14207008efd8f11f80d06e | 17,314 | null | """X18: E04 with compositional consistency loss.
Creates state-level supervision by matching T=2 intermediate states with T=1
final states when the T=2 output equals the T=1 input (string comparison only).
No arithmetic is computed - only evaluator-supplied values are compared.
"""
from __future__ import annotations
... | {
"id": "1bdecbdf-39f6-437a-9747-af2fb0cc9ce2",
"created_at": "2026-08-14 02:49:00.780924+00:00",
"db_md5": "528f8b4222adcfe6d3b2d825415ae284",
"submitter": "chad-atexpedient",
"github_login": "chad-atexpedient",
"run_id": "dfeed957-59d1-4037-be91-de9c71ca9ba7",
"tier": "easy",
"dataset_id": "e5",
"st... | {
"score": {
"mean_loss": 2.236298833848905,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.00875
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.2660368825288097,
"example_count": 600,
"... |
1be8cf73-7421-476a-b1c3-0c1667c1396c | easy | karanganesan | 2026-08-05 11:42:07.404205+00:00 | succeeded | c7ea1fbcd450c2a9b52ce079e319c13c631750fb24aa6815b9423af874d61418 | 22,081 | 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": "1be8cf73-7421-476a-b1c3-0c1667c1396c",
"created_at": "2026-08-05 11:42:07.404205+00:00",
"db_md5": "f76df226a672ed2c5e8a1a270ccdbe07",
"submitter": "Karan Ganesan",
"github_login": "karanganesan",
"run_id": "7949484a-4db0-4b80-8a3e-33e1bb3e5e5b",
"tier": "easy",
"dataset_id": "e1",
"status": ... | {
"score": {
"mean_loss": 4.010586261749268,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0033333334140479565
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 5.647897243499756,
"example_count": 100... |
1be9e852-b7af-4474-95d1-698bc37f8095 | easy | khushidahi | 2026-08-05 23:48:28.625354+00:00 | succeeded | e09187852e736f9dfe96b0f3fb78c88a9443c03b3dc634103c2947ca2a0a9f05 | 20,286 | null | """R21 hierarchical value-and-scratch 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 ta... | {
"id": "1be9e852-b7af-4474-95d1-698bc37f8095",
"created_at": "2026-08-05 23:48:28.625354+00:00",
"db_md5": "0c456391a86f20ae055846057d39b8c6",
"submitter": "khushidahi",
"github_login": "khushidahi",
"run_id": "7b650bcd-84f6-48a8-994c-f53af75a17cc",
"tier": "easy",
"dataset_id": "e4",
"status": "succ... | {
"score": {
"mean_loss": 2.0798585414886475,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.007083333563059568
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.0786077976226807,
"example_count": 12... |
1bf09c84-c990-46e2-82c3-0832384bd8d3 | easy | amirmeisami | 2026-08-15 22:46:04.192521+00:00 | succeeded | 046d897f3e1ae102dd28d3f6d6c2a31f4ed10ae3a51da666dda13943bbd6fef6 | 21,100 | null | """E011-A1: compact closed-state digit transition."""
from __future__ import annotations
import math
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
TokenLossBatch,
assert_model_state,
)
... | {
"id": "1bf09c84-c990-46e2-82c3-0832384bd8d3",
"created_at": "2026-08-15 22:46:04.192521+00:00",
"db_md5": "9f03f9fd147d8a6094caebf3a8e431b1",
"submitter": "Amir Meisami",
"github_login": "amirmeisami",
"run_id": "6be53bba-6aac-439e-9877-1ee736696b4b",
"tier": "easy",
"dataset_id": "e7",
"status": "s... | {
"score": {
"mean_loss": 1.8734543418127392,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.2648841366995252
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 1.814257764816284,
"example_count": 85,
... |
1bf47333-9708-45b7-89ad-e35899570837 | easy | erdavis0 | 2026-08-24 09:21:49.162204+00:00 | succeeded | b55593fc1495c8808062cd69554206bd2e9b09136f54bb7e28ab99edbd357bcd | 13,940 | null | """Field-aligned learned digit machine.
The fixed code in this model only exposes the public prompt grammar: delimiter
tokens select three fields and each field is right-aligned without interpreting
the value of any digit. All cross-field computation, local propagation,
workspace updates, and rendering are learned fr... | {
"id": "1bf47333-9708-45b7-89ad-e35899570837",
"created_at": "2026-08-24 09:21:49.162204+00:00",
"db_md5": "e47e69188792d13dc82454d0c65beec0",
"submitter": "Ethan",
"github_login": "erdavis0",
"run_id": "ea9554ac-d382-4545-9b2c-595dfccd5bb7",
"tier": "easy",
"dataset_id": "e4",
"status": "succeeded",... | {
"score": {
"mean_loss": 4.910223606060168,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.002916666666666667
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 4.92915417723948,
"example_count": 1200,... |
1bf8fad0-4bf8-403a-b14e-b8117c736317 | easy | karanganesan | 2026-08-08 05:23:26.235360+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": "1bf8fad0-4bf8-403a-b14e-b8117c736317",
"created_at": "2026-08-08 05:23:26.235360+00:00",
"db_md5": "3e8439ba8ab561733e7dc883fab4ecb2",
"submitter": "Karan Ganesan",
"github_login": "karanganesan",
"run_id": "27626c49-a2f9-45e5-a60e-94df7526849f",
"tier": "easy",
"dataset_id": "e5",
"status": ... | {
"score": {
"mean_loss": 7.221836805343628,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.005000000121071935
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 7.798732757568359,
"example_count": 600,... |
1bf95804-0286-4a71-b38c-83fd50dcb0ca | easy | DDanlov | 2026-08-17 18:30:44.905361+00:00 | succeeded | 6d407ea1a3414c04ed1402930d4fce8eb419820575754f0466dcca3e07062b5b | 23,097 | 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": "1bf95804-0286-4a71-b38c-83fd50dcb0ca",
"created_at": "2026-08-17 18:30:44.905361+00:00",
"db_md5": "450c76f13d256b71747385c4b8555e90",
"submitter": "DDanlov",
"github_login": "DDanlov",
"run_id": "ec791b1c-1af7-4dac-a0e9-f414088ea659",
"tier": "easy",
"dataset_id": "e1",
"status": "succeeded"... | {
"score": {
"mean_loss": 2.7066803638355257,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.08333333387970925
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.7127143242465914,
"example_count": 100... |
1bff1242-1709-4bbe-98b4-c96ea140bb3a | easy | erdavis0 | 2026-08-25 02:46:36.820372+00:00 | succeeded | 150d173c75338db2a7113c981ff2f114e97e4da3a7ca5424065633b1be6b8905 | 26,183 | null | """Reversible augmented-state transducer over aligned learned digit tapes.
Public delimiter syntax only routes opaque token categories into distinct,
right-aligned N, X, and T tapes. Learned local cells compile separate
interaction, carry, and modulus-conditioned workspaces. N is isolated from
interaction and carry ... | {
"id": "1bff1242-1709-4bbe-98b4-c96ea140bb3a",
"created_at": "2026-08-25 02:46:36.820372+00:00",
"db_md5": "fad6ce63bd264638f733d308596ed7c1",
"submitter": "Ethan",
"github_login": "erdavis0",
"run_id": "f34e31e1-807a-486e-845b-90e6c557c3db",
"tier": "easy",
"dataset_id": "e10",
"status": "succeeded"... | {
"score": {
"mean_loss": 5.806352138519287,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.09262626804411411
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 4.715620994567871,
"example_count": 125,
... |
1c01d43f-a894-4af8-8891-7730449278f2 | easy | gauravmishra | 2026-08-07 21:10:36.234731+00:00 | succeeded | cfc18b3c64b346bf94ebadb2f3b98184a4dd5b0581d252b3a312fccdb9c89a40 | 3,755 | null | """Four-pass Transformer with four independently parameterized blocks."""
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 = 12... | {
"id": "1c01d43f-a894-4af8-8891-7730449278f2",
"created_at": "2026-08-07 21:10:36.234731+00:00",
"db_md5": "32524bbcc61f82194812c606dc3efefd",
"submitter": "Gaurav Mishra",
"github_login": "gauravmishra",
"run_id": "fbee66a8-6dcd-4913-b25b-d3229efd7140",
"tier": "easy",
"dataset_id": "e5",
"status": ... | {
"score": {
"mean_loss": 2.239430902275278,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0054166666790843015
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.254176258506262,
"example_count": 600... |
1c02861f-f83d-4260-a4e8-205e434557c8 | easy | velocizapkar | 2026-08-06 22:14:54.604560+00:00 | succeeded | d1a209da9e5f94b109775a6512411af852b9c710e72c882d036c87985bd94aa3 | 55,677 | null | """Fast packed residual-free Attention-Residual transformer with Muon.
The model is a generic learned sequence model. It contains no parser,
arithmetic primitive, task constant, generated data, label feedback, or answer
construction. Every macro-state is produced by a serial pack of attention and
MLP transformations... | {
"id": "1c02861f-f83d-4260-a4e8-205e434557c8",
"created_at": "2026-08-06 22:14:54.604560+00:00",
"db_md5": "8e8a5a48165d28e36263bd64cce5f4f1",
"submitter": "Aakanksh Zarapkar",
"github_login": "velocizapkar",
"run_id": "1ffb8003-9531-4c97-9ae9-16004b3c54d4",
"tier": "easy",
"dataset_id": "e1",
"statu... | {
"score": {
"mean_loss": 2.3434978723526,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.07999999821186066
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.7521066665649414,
"example_count": 100,
... |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.