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 |
|---|---|---|---|---|---|---|---|---|---|---|
06099976-0e65-49b0-bea1-bcd87a0cdae7 | easy | yashkant | 2026-08-12 04:03:15.028713+00:00 | succeeded | 36c44133e7753e3fc0d780348e6df5b50afef09a65d62a38736e4282cab2a777 | 31,765 | null | """Round1702 three-projections-per-T Dykstra recursion on the D416 carrier.
Each parsed public-T cycle applies an A-B-A sequence through one shared learned
smooth projection, reinjecting the immutable anchor on every call and updating
the corresponding Dykstra correction memory after each projection. All active
cycle... | {
"id": "06099976-0e65-49b0-bea1-bcd87a0cdae7",
"created_at": "2026-08-12 04:03:15.028713+00:00",
"db_md5": "af932eb3600971a2c1c83c57fa57dc87",
"submitter": "Yash Kant",
"github_login": "yashkant",
"run_id": "0ba19f0d-2a26-4876-b1b7-a12ea265fba2",
"tier": "easy",
"dataset_id": "e3",
"status": "succeed... | {
"score": {
"mean_loss": 2.118978015272693,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.015000000158324839
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.119907796307677,
"example_count": 800,... |
061453e8-5023-43a6-8d55-25a1c57ab045 | easy | chuk-uzowihe | 2026-08-12 16:54:16.433477+00:00 | failed | 657a7d1ea07a5ad051827d0868fc27944cd802388bbd7381c630ad9a8f3e734a | 20,826 | 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": "061453e8-5023-43a6-8d55-25a1c57ab045",
"created_at": "2026-08-12 16:54:16.433477+00:00",
"db_md5": "719c1529bcc5a47c4827affc01e2ab75",
"submitter": "Chuk Uzowihe",
"github_login": "chuk-uzowihe",
"run_id": "d986498c-f076-4691-a547-9c96d5506351",
"tier": "easy",
"dataset_id": "e1",
"status": "... | null
|
0628cff8-fb45-451d-ac70-28540992162d | easy | CodeReclaimers | 2026-08-05 15:03:51.022783+00:00 | succeeded | 8882f29aee182512fd7027f5010c4da9ffbe2f6394fd7631e245ef7e354cf67e | 8,629 | null | """v7: v6 + squaring-symmetric transition rows and per-channel temperature.
Root-recovery fix for datasets that train only at T >= 2 (E3/E4: T=2 only):
transitions supervised only through M^T can converge to a wrong T-th root,
failing the T=1 rung. Constraint added: rows r and m-r share parameters,
encoding (m-r)^2 = ... | {
"id": "0628cff8-fb45-451d-ac70-28540992162d",
"created_at": "2026-08-05 15:03:51.022783+00:00",
"db_md5": "4753604ff69e4df2e54d8af8b66b847b",
"submitter": "CodeReclaimers",
"github_login": "CodeReclaimers",
"run_id": "60ddc6b7-d610-410c-9176-d4d1d9b028c9",
"tier": "easy",
"dataset_id": "e1",
"status... | {
"score": {
"mean_loss": 6.633669791256125e-08,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 1.0
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 8.451404198694945e-08,
"example_count": 100,
... |
062b37f8-68cc-4958-b9bc-cec1475a7104 | easy | shreyash-chonkie | 2026-08-26 13:51:51.581025+00:00 | succeeded | 3145129994429d56bf6391592faa701e97f08db0bf9ec43e1b2e4b64269b5eec | 10,856 | null | """Compact dual-state recursive reasoner with detached outer credit paths."""
from __future__ import annotations
import math
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
TokenLossBatch,
... | {
"id": "062b37f8-68cc-4958-b9bc-cec1475a7104",
"created_at": "2026-08-26 13:51:51.581025+00:00",
"db_md5": "75d84055af6e2f05872692bbe998d545",
"submitter": "Shreyash",
"github_login": "shreyash-chonkie",
"run_id": "d9586575-39bb-4f3a-9820-0829fe614daa",
"tier": "easy",
"dataset_id": "e1",
"status": "... | {
"score": {
"mean_loss": 2.879069447517395,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.031666665337979794
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.665745258331299,
"example_count": 100,... |
062be5f4-91d2-4cf5-b42c-e6497f4ba6f6 | easy | KaustubhKumar05 | 2026-08-25 16:35:34.768835+00:00 | succeeded | 03d0c373e594aa5715323fd2ba901ecc5a0250f4deb9d9e32371e95dc8a36c27 | 9,417 | null | """abacus: faithful port of the Abacus recipe (McLeish et al., arXiv:2405.17399).
The load-bearing idea is the embedding. Each digit is indexed by its
SIGNIFICANCE WITHIN ITS OWN NUMBER (0 = least-significant), from a table
SHARED across N, x and T. Digit k of N and digit k of x therefore land on the
same embedding an... | {
"id": "062be5f4-91d2-4cf5-b42c-e6497f4ba6f6",
"created_at": "2026-08-25 16:35:34.768835+00:00",
"db_md5": "80b03c6db927f2e12cf645f2453373df",
"submitter": "koz",
"github_login": "KaustubhKumar05",
"run_id": "446c4360-44b5-4c85-9b2c-08af4171dfae",
"tier": "easy",
"dataset_id": "e6",
"status": "succee... | {
"score": {
"mean_loss": 6.266285419464111,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.03513513784855604
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 6.44103479385376,
"example_count": 60,
... |
062bfec5-a42d-44fe-bfc3-a9c32983ffbe | easy | pculaf | 2026-08-25 07:36:03.724608+00:00 | succeeded | 3cbc7a7efb706619bf0fd6603b5f97aa1da66c46dca5a6a8698152075f45d559 | 7,754 | null | """Small single-pass post-RMSNorm Transformer for One Layer Deeper."""
from __future__ import annotations
from collections.abc import Iterable
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
... | {
"id": "062bfec5-a42d-44fe-bfc3-a9c32983ffbe",
"created_at": "2026-08-25 07:36:03.724608+00:00",
"db_md5": "90672240fbe1c35ef701652734edcf8b",
"submitter": "Pavle Culafic",
"github_login": "pculaf",
"run_id": "41f201fd-62e3-4ac4-a9ef-2b2d054cb6a1",
"tier": "easy",
"dataset_id": "e6",
"status": "succe... | {
"score": {
"mean_loss": 1.9887205958366394,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.05698198452591896
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 1.9936890602111816,
"example_count": 60,... |
062dae80-398c-47ff-a2a1-b67f54898662 | easy | DDanlov | 2026-08-10 10:13:09.880745+00:00 | succeeded | 54df5fb02137eeadeb2bd3d137cce1e133a30d32a4534ce2179a3412f6a090c9 | 16,615 | null | from __future__ import annotations
import sys
import os
import math
import time
import torch
import torch.nn as nn
import torch.nn.functional as F
from torch.optim.optimizer import Optimizer
VOCAB_SIZE = 17
DIGIT_OFFSET = 7
def init_cutting_the_skip(module):
"""SVD-based orthogonal initialization to ensure initia... | {
"id": "062dae80-398c-47ff-a2a1-b67f54898662",
"created_at": "2026-08-10 10:13:09.880745+00:00",
"db_md5": "4b3b91b45de5218228dd0cee2acc873c",
"submitter": "DDanlov",
"github_login": "DDanlov",
"run_id": "d6efdbae-ff27-47bb-99e7-42d2492988c0",
"tier": "easy",
"dataset_id": "e1",
"status": "succeeded"... | {
"score": {
"mean_loss": 4.527745164769147,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.006666666666666667
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 4.7633488729818545,
"example_count": 100... |
062fc282-0096-4fe5-ab07-f44767870af3 | easy | DDanlov | 2026-08-18 00:14:57.910753+00:00 | succeeded | 115d46ec00ec18546afcf1ba31d74ac788944592739c4e1a211477fa256dc823 | 23,102 | null | """
Official Submission for One Layer Deeper Challenge
Architecture: 4-Layer Recurrent Core Pure No-Norm No-Skip DKS Reasoning Model:
- 4 Distinct Pure No-Norm No-Skip Transformer Layers in the Recurrent Core: L1 -> L2 -> L3 -> L4
- Unrolled T times for serial modular squaring
- dim=512, 16 heads, d_ff=1024, RohanShamp... | {
"id": "062fc282-0096-4fe5-ab07-f44767870af3",
"created_at": "2026-08-18 00:14:57.910753+00:00",
"db_md5": "ef8f2b4e45393e17f33e8d6b42c30e5d",
"submitter": "DDanlov",
"github_login": "DDanlov",
"run_id": "69d3fa1e-4a37-48c2-8692-33ba2fa46cdd",
"tier": "easy",
"dataset_id": "e1",
"status": "succeeded"... | {
"score": {
"mean_loss": 2.016082110334686,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.02833333346992731
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 1.9776930132908608,
"example_count": 100,... |
0632669b-7087-4b91-ba4f-513b781913af | easy | khushidahi | 2026-08-02 06:23:36.024480+00:00 | succeeded | 8ae552d46a425d6d557e4206bfb198698e761a3eebb0e6cc6566fe349825cedb | 9,304 | null | """Prelude -> recalled shared core -> coda model for One Layer Deeper.
R2 architecture reset:
- use a real one-block prelude and one-block coda;
- apply one shared recurrent Transformer core four times;
- inject immutable prelude context at every recurrent update;
- use sandwich RMS normalization for stable repeated a... | {
"id": "0632669b-7087-4b91-ba4f-513b781913af",
"created_at": "2026-08-02 06:23:36.024480+00:00",
"db_md5": "f66b3203cbfbb8b14b183bbac3a7ab1e",
"submitter": "khushidahi",
"github_login": "khushidahi",
"run_id": "64577a04-ac76-43a6-89d3-9db9a9af1110",
"tier": "easy",
"dataset_id": "e5",
"status": "succ... | {
"score": {
"mean_loss": 2.1740094423294067,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.008750000037252903
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.1808724403381348,
"example_count": 60... |
063a5c38-12cf-4205-958c-566742d3d27c | easy | DDanlov | 2026-08-11 19:32:07.938029+00:00 | failed | f4593f57df4658db88b1a37a6f891fd203f4241044a357f878bd6cfb535f6fb9 | 19,083 | 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 bias-correcte... | {
"id": "063a5c38-12cf-4205-958c-566742d3d27c",
"created_at": "2026-08-11 19:32:07.938029+00:00",
"db_md5": "82a7bf741a1bd9eace5c79253db302a7",
"submitter": "DDanlov",
"github_login": "DDanlov",
"run_id": "8296cffe-ff3a-424c-803a-24f05ef8419a",
"tier": "easy",
"dataset_id": "e1",
"status": "failed",
... | null
|
063c6883-dac7-46e6-9fb1-94a514805b99 | easy | DDanlov | 2026-08-11 21:18:59.599525+00:00 | succeeded | 6f755f5eb6bc8d770a8d317f704a799070752a4846c0d9a3a7fd99ec205ddb1a | 19,597 | null | from __future__ import annotations
import sys
import os
import math
import time
import torch
import torch.nn as nn
import torch.nn.functional as F
from torch.optim.optimizer import Optimizer
VOCAB_SIZE = 17
DIGIT_OFFSET = 7
def init_cutting_the_skip(module):
"""SVD-based orthogonal initialization to ensure initia... | {
"id": "063c6883-dac7-46e6-9fb1-94a514805b99",
"created_at": "2026-08-11 21:18:59.599525+00:00",
"db_md5": "e42e0147690268dcfd5b91d0f981d798",
"submitter": "DDanlov",
"github_login": "DDanlov",
"run_id": "59d42188-99bf-4a1b-8b44-fcf8cdce8b37",
"tier": "easy",
"dataset_id": "e5",
"status": "succeeded"... | {
"score": {
"mean_loss": 30.239296751845046,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0012500000124176342
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 32.98478314587888,
"example_count": 60... |
06429e6a-0886-4cd7-a113-5861725d5127 | easy | briheet | 2026-08-05 12:09:42.345340+00:00 | succeeded | 7710fcada2948bfd0e997def3b81ab1c3086c06b651fb66a65c60a1ecb363dc0 | 6,380 | null | """Basic single-pass Transformer with PyTorch AdamW."""
from __future__ import annotations
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
assert_model_state,
)
D_MODEL = 96
NUM_HEADS = 4
LI... | {
"id": "06429e6a-0886-4cd7-a113-5861725d5127",
"created_at": "2026-08-05 12:09:42.345340+00:00",
"db_md5": "d507e501b89d6e6f60a9fb0e188a4e62",
"submitter": "Aizenn",
"github_login": "briheet",
"run_id": "5609bb64-58cb-4616-a5ac-4775306107d1",
"tier": "easy",
"dataset_id": "e1",
"status": "succeeded",... | {
"score": {
"mean_loss": 4.992542266845703,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.04333333298563957
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 6.1626691818237305,
"example_count": 100,... |
0645ec84-4d6e-4115-bba1-47bda639c4fb | easy | shirvani-jr | 2026-08-28 00:50:57.199549+00:00 | succeeded | 396fee419635addaf93a12d3a1446ceda5cdae5e6fca6dbe436e0fa97b7929b8 | 48,549 | 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": "0645ec84-4d6e-4115-bba1-47bda639c4fb",
"created_at": "2026-08-28 00:50:57.199549+00:00",
"db_md5": "6eecdcb811fc7c9142f7f4a70d9f9fa1",
"submitter": "Ali",
"github_login": "shirvani-jr",
"run_id": "94ef72c4-878a-4e6c-addb-7bee5108f636",
"tier": "easy",
"dataset_id": "e4",
"status": "succeeded"... | {
"score": {
"mean_loss": 2.2564433449899686,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.00375
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.2580896040427487,
"example_count": 1200,
... |
06493715-6b10-49f7-815c-6e92c52fa9c9 | easy | yashkant | 2026-08-17 00:03:50.209953+00:00 | succeeded | ddf18dba92baba7a052c5e85a979c9998522b1bc2c9131bf6b001f882f471abb | 44,697 | null | """Round2088 official-Hard carrier with a four-digit-N residual.
The exact Round1933 D273 sparse-SAM8 LR0.0075 carrier is retained. A
width-32 bounded residual is appended after every parent tensor and is
active only when public N has at least four decimal digits. Candidate
and equal-work control eagerly compute ide... | {
"id": "06493715-6b10-49f7-815c-6e92c52fa9c9",
"created_at": "2026-08-17 00:03:50.209953+00:00",
"db_md5": "f62e692f3e5a76c2d54093d4a75ba2ff",
"submitter": "Yash Kant",
"github_login": "yashkant",
"run_id": "77072b8e-81d8-4ea9-840a-a2f0311eac4f",
"tier": "easy",
"dataset_id": "e1",
"status": "succeed... | {
"score": {
"mean_loss": 2.786521077156067,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.046666666865348816
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.7859106063842773,
"example_count": 100... |
0649ecf0-7d10-4b35-b6f9-b8e13a36fce4 | easy | viridale | 2026-08-24 19:12:14.283626+00:00 | succeeded | e6cee62b2120a2e3e4e3b2b076e825c9bbc6d2a07017a2ad74c7cbe4184c8f5c | 25,377 | null | """Quadratic-coefficient Transformer with exhaustive exact support.
hypothesis: the hidden one-step edit is a small signed quadratic, but v473's single
floating scalar both lacked a Transformer controller and made its support test in
an approximate phase carrier. A right-aligned Abacus Transformer should acqui... | {
"id": "0649ecf0-7d10-4b35-b6f9-b8e13a36fce4",
"created_at": "2026-08-24 19:12:14.283626+00:00",
"db_md5": "5d1d212da4c1ba94fdbfa01df5118a79",
"submitter": "priormancer",
"github_login": "viridale",
"run_id": "0d61e105-30b5-4bd9-a527-ff1bf27ead7f",
"tier": "easy",
"dataset_id": "e6",
"status": "succe... | {
"score": {
"mean_loss": 0.0,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 1.0
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 0.0,
"example_count": 60,
"exact_accuracy": 1.0,
"c... |
064cbf80-b15d-445b-a145-13819903104e | easy | oupadhyay | 2026-08-10 16:58:19.012340+00:00 | succeeded | c299ba6f22537b3a59beec5d17bf47711e4167717b5ce570342fec1367f4102d | 8,817 | null | """Identity-initialized dynamic-depth learned transition."""
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_MARKER, X_MARKER, T_MARKER, ANSWER_MARKER,... | {
"id": "064cbf80-b15d-445b-a145-13819903104e",
"created_at": "2026-08-10 16:58:19.012340+00:00",
"db_md5": "144f0c53486bdb02276266d6fd61d88a",
"submitter": "Ojasw Upadhyay",
"github_login": "oupadhyay",
"run_id": "c158c2a5-d758-4c03-a449-77b2c2c1e7cd",
"tier": "easy",
"dataset_id": "e5",
"status": "s... | {
"score": {
"mean_loss": 2.192411266804278,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0025
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.183901071548462,
"example_count": 600,
"ex... |
06586bb5-b4f6-495d-927c-f07fc557df30 | easy | shirvani-jr | 2026-08-30 19:56:56.770818+00:00 | succeeded | c55a9b3b218530cd0d9e679d26165b89d5dfe4a8fa0744f2e2a8901af7748ddc | 22,841 | null | from __future__ import annotations
import math
import time
from contextlib import nullcontext
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import ModelSpec, OptimizerBundle, OptimizerSpec, Submission, TokenLossBatch, assert_model_state
MODEL_WIDTH = 128
ATTENTION_HEADS =... | {
"id": "06586bb5-b4f6-495d-927c-f07fc557df30",
"created_at": "2026-08-30 19:56:56.770818+00:00",
"db_md5": "5371dc4989bdb508199773c68e4d2e71",
"submitter": "Ali",
"github_login": "shirvani-jr",
"run_id": "5ac3cebe-e0b0-4175-bd89-ef3c50b98f10",
"tier": "easy",
"dataset_id": "e4",
"status": "succeeded"... | {
"score": {
"mean_loss": 7.608413326350065,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.008854166681412607
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 7.546441689022275,
"example_count": 1200... |
065eb2ef-5c26-432d-a15e-8db1e106e134 | easy | sapient-sapiens | 2026-08-21 21:30:19.721583+00:00 | succeeded | edcf0838907fd61eae4f4364b1aa866f12efea83c2437ad4801c71620f9bab91 | 33,481 | 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": "065eb2ef-5c26-432d-a15e-8db1e106e134",
"created_at": "2026-08-21 21:30:19.721583+00:00",
"db_md5": "7d113c64064d392078ef970876e0b928",
"submitter": "Ertondy",
"github_login": "sapient-sapiens",
"run_id": "18e63238-50bd-46d0-8ac5-ad676957c62c",
"tier": "easy",
"dataset_id": "e9",
"status": "su... | {
"score": {
"mean_loss": 4.495325419404344,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.06000000203649203
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 4.755157947540283,
"example_count": 90,
... |
066a2c12-0d4a-48ba-a054-ddf0ccdcf9d3 | easy | rougee | 2026-08-25 00:35:09.383700+00:00 | succeeded | db6e41e65004d7fd599951d5654d7fab676fefec7d7c275d1c3ce9cd8938ebda | 12,114 | null | """Official Submittable Champion for One Layer Deeper (Reinjected Quadratic-Cayley Core).
Proven, Certified Architecture:
- Core Type: Dynamic Cayley Orthogonal Recurrent Core (Strictly Norm-Preserving |λ_i| ≡ 1)
- Algebraic Non-Linearity: Quadratic Bilinear Interaction (ax^2 + bx + c inductive bias)
- Topology: d_mod... | {
"id": "066a2c12-0d4a-48ba-a054-ddf0ccdcf9d3",
"created_at": "2026-08-25 00:35:09.383700+00:00",
"db_md5": "0abc7bf11914052483308c0c12ae2925",
"submitter": "Rougee",
"github_login": "rougee",
"run_id": "03ea8f5e-9495-4fec-b475-a67b7449779c",
"tier": "easy",
"dataset_id": "e1",
"status": "succeeded",
... | {
"score": {
"mean_loss": 10.507057666778564,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.03999999910593033
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 11.98444938659668,
"example_count": 100,... |
0673c842-a6b8-4ffa-8248-d29579785557 | easy | erdavis0 | 2026-08-21 00:35:02.407567+00:00 | failed | 2c07dc032489f42a8f114f14274a9dec222dbf8778eb5526de91ff8258c00f1e | 34,915 | null | """Sealed neural local rewrite machine with sparse center-write events."""
from __future__ import annotations
import math
import time
from dataclasses import dataclass
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerS... | {
"id": "0673c842-a6b8-4ffa-8248-d29579785557",
"created_at": "2026-08-21 00:35:02.407567+00:00",
"db_md5": "e9b2ff7274dd5e2c75c2c3d0842b9370",
"submitter": "Ethan",
"github_login": "erdavis0",
"run_id": "12357474-2eea-477a-bbde-7629e1936dc3",
"tier": "easy",
"dataset_id": "e3",
"status": "failed",
... | null
|
067569d6-7228-4d13-83d9-7bd8c3ed66a1 | easy | jordanrubin | 2026-08-01 20:13:06.261884+00:00 | succeeded | 2c6adbc91db77091eaa1fdb53b2b15fdc6e05306608a2440be5e6a2aa2591b8a | 18,183 | null | """Looped bidirectional selective-SSM (Mamba-style), pure PyTorch.
The statefulness bet: input-dependent SSM dynamics along the token axis,
with the block weight-tied and applied NUM_LOOPS times in depth with input
injection — recurrence in depth is where the task's T-iteration lives.
The selective scan is computed i... | {
"id": "067569d6-7228-4d13-83d9-7bd8c3ed66a1",
"created_at": "2026-08-01 20:13:06.261884+00:00",
"db_md5": "35fec302dbad6ab1de27b5092ede0b6a",
"submitter": "Jordan Rubin",
"github_login": "jordanrubin",
"run_id": "1a1a24b8-07fe-4995-bb31-340630aa4646",
"tier": "easy",
"dataset_id": "e3",
"status": "s... | {
"score": {
"mean_loss": 5.128315362484105,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.00937499979045242
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 4.457168238675539,
"example_count": 800,
... |
0676a34a-2359-40de-88d2-271cdfa5545a | easy | nikolageorgiev2000 | 2026-08-24 02:02:31.984654+00:00 | succeeded | d721bd60a0846b23bab33edc6cc91fcd4dca4533e9885d2814b0b44f0152a160 | 7,833 | null | """Atomic T1 recurrent modular squaring with Neural-GPU CGRUs."""
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
DIGIT_OFFSET = 7
DIGITS = 10
C... | {
"id": "0676a34a-2359-40de-88d2-271cdfa5545a",
"created_at": "2026-08-24 02:02:31.984654+00:00",
"db_md5": "7299ce0d2fe41552461d0aacc60de463",
"submitter": "Nikola Georgiev",
"github_login": "nikolageorgiev2000",
"run_id": "03470287-6c4e-475e-82f8-955a59fedf4e",
"tier": "easy",
"dataset_id": "e1",
"s... | {
"score": {
"mean_loss": 9.874643802642822,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.036666667088866234
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 10.816900253295898,
"example_count": 100... |
067c968b-fdd2-457c-ad86-5c557614252d | easy | DDanlov | 2026-08-15 21:02:44.288806+00:00 | succeeded | 3eb517cb9667863720ce75d439fd67ebc36b79a26534e0b145dad12284313550 | 17,950 | null | """
Official Submission for One Layer Deeper Challenge
Architecture: Deep Equilibrium Port-Hamiltonian Deliberation Network (Model C v3 / DEPH-Net)
- Input: Multi-Scale Linear Harmonic Torus Frequency Encoder (omega_j = 1..64)
- Recurrent Core: Continuous Port-Hamiltonian Dynamical Flow dz/dt = (J - R) * grad_E
* J =... | {
"id": "067c968b-fdd2-457c-ad86-5c557614252d",
"created_at": "2026-08-15 21:02:44.288806+00:00",
"db_md5": "623970f9d216ffb45b2ef4f8507de941",
"submitter": "DDanlov",
"github_login": "DDanlov",
"run_id": "d4b7c8e5-6e15-4dcc-ba1a-58f994169fb7",
"tier": "easy",
"dataset_id": "e5",
"status": "succeeded"... | {
"score": {
"mean_loss": 2.4435929158256338,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0033333333581686024
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.4521222737337975,
"example_count": 6... |
068238b8-82aa-40e7-88de-5bb08278cce9 | easy | DDanlov | 2026-08-08 11:36:48.881017+00:00 | succeeded | 0e9c1c3fdd85d0e767158fd9555649ae1ec9927d0677e0459ebafb4c0a4f8779 | 16,048 | 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": "068238b8-82aa-40e7-88de-5bb08278cce9",
"created_at": "2026-08-08 11:36:48.881017+00:00",
"db_md5": "6896be6b7bcdaaee583f5cd47073535c",
"submitter": "DDanlov",
"github_login": "DDanlov",
"run_id": "b3ad5cea-a42a-47df-85c5-a5244b676195",
"tier": "easy",
"dataset_id": "e5",
"status": "succeeded"... | {
"score": {
"mean_loss": 44.58081929307704,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0125
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 45.825776386688645,
"example_count": 600,
"e... |
0689ebbe-50d8-444b-a5bb-6133bfcf5257 | easy | khushidahi | 2026-08-20 02:23:03.591614+00:00 | succeeded | 955f5ed5f89010c77013bba201b55c99eaa7b51c1e55fbd47a834b48ed463719 | 38,087 | null | """R280: tied-codebook semantic quadratic-root carry/reduce machine.
This candidate targets functional-root recovery when the evaluator supplies
only composed endpoints such as T=2. It trains only from evaluator-owned final
labels and always executes the supplied recurrence depth. The model never
converts N or X to ... | {
"id": "0689ebbe-50d8-444b-a5bb-6133bfcf5257",
"created_at": "2026-08-20 02:23:03.591614+00:00",
"db_md5": "1002191406bcfcc13c3235afb43ab39b",
"submitter": "khushidahi",
"github_login": "khushidahi",
"run_id": "f04df0d5-3b1f-4669-ba2f-cd5eb8a19a80",
"tier": "easy",
"dataset_id": "e3",
"status": "succ... | {
"score": {
"mean_loss": 2.2182881377857617,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.010625000013969838
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.203656084757168,
"example_count": 800... |
068c23b3-e3c4-4d52-9cf3-f5257bf4a5e4 | easy | viridale | 2026-08-15 07:20:05.227547+00:00 | succeeded | 8fa1edffff511f1c49d09cc29e307410919ad3b9cba1ad14bc1378d58ee70032 | 20,273 | null | """Independent base-aligned character supervision on v633.
hypothesis: a bounded injective FoNE representation of each supplied complete answer
may provide smoother whole-string credit than token CE while leaving the learned
transition and every prediction logit unchanged.
axis: loss
target: P142 matched E6 co... | {
"id": "068c23b3-e3c4-4d52-9cf3-f5257bf4a5e4",
"created_at": "2026-08-15 07:20:05.227547+00:00",
"db_md5": "1851b4e186e97261341271f1ca466184",
"submitter": "priormancer",
"github_login": "viridale",
"run_id": "0f8c9121-f0c8-415c-b883-76e34ad374d8",
"tier": "easy",
"dataset_id": "e6",
"status": "succe... | {
"score": {
"mean_loss": 2.401442050933838,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.3587838113307953
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.062329053878784,
"example_count": 60,
... |
0690ba0d-39ac-445d-b1f4-2a089912daef | easy | AFeinfield | 2026-08-14 17:54:34.325971+00:00 | succeeded | 0a8d1c270bb99ff74aa6af850d22a5ee7c9d8a7aab40b84069e9b012807de846 | 9,209 | null | """Looped (weight-tied) bidirectional Transformer, d=256, 8 iterations.
038: T-weighted loss — model passes parsed per-example T through aux;
token_training_loss weights CE by 1/T (low-T emphasis = the ranking metric).
to amortize the evaluator's ~200ms/iteration overhead into 8x more steps.
prior. Same parameter coun... | {
"id": "0690ba0d-39ac-445d-b1f4-2a089912daef",
"created_at": "2026-08-14 17:54:34.325971+00:00",
"db_md5": "02a17ea51e887b7145193eaea487a700",
"submitter": "AFeinfield",
"github_login": "AFeinfield",
"run_id": "e10071e7-720b-4489-9429-fbf9dfa870c5",
"tier": "easy",
"dataset_id": "e5",
"status": "succ... | {
"score": {
"mean_loss": 8.33051099215212,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.006666666666666666
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 8.552181915317416,
"example_count": 600,
... |
06939849-ad8a-45d3-b347-14c3f0061ef2 | easy | sapient-sapiens | 2026-08-14 18:19:40.780985+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": "06939849-ad8a-45d3-b347-14c3f0061ef2",
"created_at": "2026-08-14 18:19:40.780985+00:00",
"db_md5": "1efadda9e01ae84ec2857432a666c542",
"submitter": "Ertondy",
"github_login": "sapient-sapiens",
"run_id": "efcbd439-ec32-44db-bb17-ebc78484a7d6",
"tier": "easy",
"dataset_id": "e2",
"status": "su... | {
"score": {
"mean_loss": 2.3695200253860094,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0033333333333333335
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.3394377192830653,
"example_count": 3... |
06c05f8c-5de8-4104-a6e4-a3377c15f3d5 | easy | erdavis0 | 2026-08-27 01:39:08.581223+00:00 | succeeded | 2631fff557194c25861f5cbd2a3339df5a8a8980d811bc985ebcaec80ee4bdc0 | 31,299 | null | """Typed-prefix reversible transition B5: CE/sequence-mixture hybrid.
Public delimiter syntax routes opaque N, X, and T fields into three strict
functional roles. X initializes only an evolving auxiliary tape, N initializes
only an invariant environment lane, and T reaches only a learned monotone
prefix controller. ... | {
"id": "06c05f8c-5de8-4104-a6e4-a3377c15f3d5",
"created_at": "2026-08-27 01:39:08.581223+00:00",
"db_md5": "2981fa4ef2c40f38bd0d3090c6d64a88",
"submitter": "Ethan",
"github_login": "erdavis0",
"run_id": "c314c024-86b9-4ede-a878-04dc80dd5b6e",
"tier": "easy",
"dataset_id": "e3",
"status": "succeeded",... | {
"score": {
"mean_loss": 3.696906816706477,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.013125000023283064
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 3.680556285065365,
"example_count": 800,... |
06c7373c-b903-49cc-ad2d-d2f8d307e00f | easy | nikolageorgiev2000 | 2026-08-04 09:42:26.024900+00:00 | succeeded | fe4c052cf2410ed697f85a9fe1a83bd491557858255b038412af6fd79fe09c6e | 8,615 | null | """T-composed categorical-digit reasoner with fully learned pair/reduction maps."""
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,
asser... | {
"id": "06c7373c-b903-49cc-ad2d-d2f8d307e00f",
"created_at": "2026-08-04 09:42:26.024900+00:00",
"db_md5": "1ef5988b8f2d3469a7e0a8c2c679c946",
"submitter": "Nikola Georgiev",
"github_login": "nikolageorgiev2000",
"run_id": "f3701707-c230-49d5-b8a2-d764a01231e1",
"tier": "easy",
"dataset_id": "e3",
"s... | {
"score": {
"mean_loss": 8.614283159123861,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.018750000009313227
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 8.237783914661527,
"example_count": 800,... |
06c80902-5e0c-4dc4-9177-593be6018ff7 | easy | DDanlov | 2026-08-27 22:18:25.597907+00:00 | succeeded | 044d7228beacd6019e6bbb92bc89fcccc16fb85374b05c7230ba36d15aa351e6 | 39,714 | null | import math
import time
from dataclasses import dataclass
from typing import Optional, Tuple, Dict, Any, List, Union
import torch
import torch.nn as nn
import torch.nn.functional as F
from torch.optim import Optimizer
try:
from onelayerdeeper.api import Submission, OptimizerBundle, ModelSpec, OptimizerSpec
except ... | {
"id": "06c80902-5e0c-4dc4-9177-593be6018ff7",
"created_at": "2026-08-27 22:18:25.597907+00:00",
"db_md5": "fec9978ae244892abd7b27bd882317fd",
"submitter": "DDanlov",
"github_login": "DDanlov",
"run_id": "ae13d653-9fd5-4601-b981-d3b4f949d82e",
"tier": "easy",
"dataset_id": "e5",
"status": "succeeded"... | {
"score": {
"mean_loss": 2.7417051792144775,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.004166666651144624
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.743121862411499,
"example_count": 600... |
06c8e265-c37a-408d-9106-777e95cc9073 | easy | sapient-sapiens | 2026-08-18 19:38:41.409186+00:00 | succeeded | e7cd0a0966b25aa0ad37f0f7443b400b6bafb76ecfb3e1c5a451f7f555c7319c | 31,705 | 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": "06c8e265-c37a-408d-9106-777e95cc9073",
"created_at": "2026-08-18 19:38:41.409186+00:00",
"db_md5": "a64dcfd8f9600605f4baa684b26c7ad3",
"submitter": "Ertondy",
"github_login": "sapient-sapiens",
"run_id": "96d37a70-076c-4ebd-a6fe-b22d71c7996e",
"tier": "easy",
"dataset_id": "e5",
"status": "su... | {
"score": {
"mean_loss": 2.675164828034142,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.005833333333333334
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.6836642887560243,
"example_count": 600... |
06cccf1b-e058-4fa3-8c59-68be582fc62a | easy | sapient-sapiens | 2026-08-13 20:26:03.081240+00:00 | failed | 3b4946fc0bc7b2d71d3adbd03543b85e62dd782e38d6c93b30829da6a66f3d20 | 9,553 | null | """E3 fixed-T=2: differentiable, field-aligned digit-state recurrence.
The model reads marker positions to extract and right-align fields, but never
converts digit tokens into numbers. A learned four-block operator is applied
twice. The first application's categorical distribution is softly re-embedded
so final-answ... | {
"id": "06cccf1b-e058-4fa3-8c59-68be582fc62a",
"created_at": "2026-08-13 20:26:03.081240+00:00",
"db_md5": "573f59d91ed7b1659dbd92d936e26d95",
"submitter": "Ertondy",
"github_login": "sapient-sapiens",
"run_id": "d5f68e15-e7de-4b66-90c7-4e2df05095d0",
"tier": "easy",
"dataset_id": "e3",
"status": "fa... | null
|
06cfb0e9-7ab5-416c-a8c5-23b0e9138cf8 | easy | DDanlov | 2026-08-26 19:19:02.374513+00:00 | failed | fbb426fa66ab4ac63f6b9ea819988cce7f6395e4dd74163e14366beae9545cc3 | 33,048 | null | import math
import time
from typing import Optional, Tuple, Dict, Any, List, Union
import torch
import torch.nn as nn
import torch.nn.functional as F
from torch.optim import Optimizer
try:
from onelayerdeeper.api import Submission, OptimizerBundle, ModelSpec, OptimizerSpec
except ImportError:
try:
from... | {
"id": "06cfb0e9-7ab5-416c-a8c5-23b0e9138cf8",
"created_at": "2026-08-26 19:19:02.374513+00:00",
"db_md5": "1951368f3383f02f577426ee3a5b70ee",
"submitter": "DDanlov",
"github_login": "DDanlov",
"run_id": "06a61a2a-2392-469d-a257-c4891f6bad78",
"tier": "easy",
"dataset_id": "e5",
"status": "failed",
... | null
|
06de3ebe-54cb-4fef-8684-0c2879f369e4 | easy | blake-camp-surge | 2026-08-04 02:12:15.414688+00:00 | succeeded | e83725104e61c303ce3eedce728401a95f7cd04ecf7651e2934b654ce16ff0e0 | 10,021 | null | """Scalar virtual neurons driven by one shared QKV-generating DAN."""
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... | {
"id": "06de3ebe-54cb-4fef-8684-0c2879f369e4",
"created_at": "2026-08-04 02:12:15.414688+00:00",
"db_md5": "ffc36182b7f0ebcd6382d983d7933049",
"submitter": "Blake Camp",
"github_login": "blake-camp-surge",
"run_id": "f4bd8b24-e334-498d-a48e-9e7f37d35945",
"tier": "easy",
"dataset_id": "e5",
"status":... | {
"score": {
"mean_loss": 2.183519983445255,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0062500000372529025
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.1845551285508504,
"example_count": 60... |
06e39d31-372a-410f-a231-6823c962b253 | easy | chad-atexpedient | 2026-08-09 21:47:32.130661+00:00 | succeeded | 9d634d4fc4c34785737c57059f42d56bcf40e3ef9c074c3917b40a38dcc5b729 | 10,010 | null | """BL7A: step-400 LR 2e-4."""
from __future__ import annotations
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
assert_model_state,
)
LANES = 6
IMMUTABLE_LANES = 3
MUTABL... | {
"id": "06e39d31-372a-410f-a231-6823c962b253",
"created_at": "2026-08-09 21:47:32.130661+00:00",
"db_md5": "3537fb8e0b44a98f537b5391d7f64b47",
"submitter": "chad-atexpedient",
"github_login": "chad-atexpedient",
"run_id": "7914e246-e64f-4c33-9f28-4e028097b7b6",
"tier": "easy",
"dataset_id": "e5",
"st... | {
"score": {
"mean_loss": 3.2345040043195086,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.004166666679084301
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 3.201442241668701,
"example_count": 600... |
06e46529-bb8b-42ab-b768-ecf8cd1e3b33 | easy | viridale | 2026-08-02 13:50:03.813551+00:00 | succeeded | 914f97d9c109315faea178caae5d79b4f92b1ffd685b12e3c140c5c491781cd2 | 52,344 | null | """Row-subset streamed five-monomial closure plus open table.
hypothesis: four ordinary rows uniquely identify the complete five-term
degree-5 coefficient tuple. Score the 12.15M new atoms on exactly four
provided rows during the one full-information update, while v102 scores
all lower families on the ful... | {
"id": "06e46529-bb8b-42ab-b768-ecf8cd1e3b33",
"created_at": "2026-08-02 13:50:03.813551+00:00",
"db_md5": "6f94add419826c62ee948d23961025f6",
"submitter": "priormancer",
"github_login": "viridale",
"run_id": "312e3866-040a-464b-9d8d-d5a4acc4c389",
"tier": "easy",
"dataset_id": "e3",
"status": "succe... | {
"score": {
"mean_loss": 0.003014546214626436,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 1.0
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 0.00301454615599774,
"example_count": 800,
"... |
06ebde04-8388-45ad-a42b-180d65b43c89 | easy | erdavis0 | 2026-08-19 09:28:01.995015+00:00 | succeeded | e6d75591b123483c27b5a4344aa991e325184d79f5ed87c2a9fa7d7390f5f5bd | 10,552 | null | """Generic learned-slot transducer with a shared low-rank latent operator."""
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,
TokenL... | {
"id": "06ebde04-8388-45ad-a42b-180d65b43c89",
"created_at": "2026-08-19 09:28:01.995015+00:00",
"db_md5": "fb39c040e0caf73a4ee6e4c2f9f4d564",
"submitter": "Ethan",
"github_login": "erdavis0",
"run_id": "b0f0c3eb-3369-41f7-93cb-777370b942e8",
"tier": "easy",
"dataset_id": "e9",
"status": "succeeded",... | {
"score": {
"mean_loss": 1.9520135521888733,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.035555556416511536
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 1.9615520238876343,
"example_count": 90... |
06ebe59d-6fb8-488c-8f18-aa69a133d038 | easy | ehonig | 2026-08-20 08:03:14.204598+00:00 | succeeded | b35a7609c5daafd376fed56f8cabe9a8dafd16f084fb2a3894b26027d1300de5 | 23,040 | 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": "06ebe59d-6fb8-488c-8f18-aa69a133d038",
"created_at": "2026-08-20 08:03:14.204598+00:00",
"db_md5": "6cea0a0bfc2945363f2ce4b882068fb8",
"submitter": "Edouardo Honig",
"github_login": "ehonig",
"run_id": "4508a67b-d918-4529-afbe-4780ad275c8d",
"tier": "easy",
"dataset_id": "e2",
"status": "succ... | {
"score": {
"mean_loss": 2.2249408960342407,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.018125000409781933
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.2204248905181885,
"example_count": 30... |
06f3f3bc-0505-4103-b44d-e01740bc34fa | easy | richardcepka | 2026-08-17 19:55:35.507323+00:00 | succeeded | 01a9c850762c5db82b090007d1c6b09538ab1e4219c213400910141b5e9ea4e8 | 43,768 | 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": "06f3f3bc-0505-4103-b44d-e01740bc34fa",
"created_at": "2026-08-17 19:55:35.507323+00:00",
"db_md5": "59ac9443ce1eeb6d4ec82c897ca38ec9",
"submitter": "Richard Cepka",
"github_login": "richardcepka",
"run_id": "5fdc6ef7-0c8d-4c67-bef4-dd5ed958ef18",
"tier": "easy",
"dataset_id": "e5",
"status": ... | {
"score": {
"mean_loss": 2.1447222232818604,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0033333332976326346
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.1465256214141846,
"example_count": 6... |
06f6ae19-eaca-4d45-851f-645340ae2d77 | easy | TZZheng | 2026-08-06 17:52:02.436651+00:00 | succeeded | 9f79ba30f8f7bdf5092d1d407abdbd3e46a560128932a685ba5742157c78632a | 7,305 | 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": "06f6ae19-eaca-4d45-851f-645340ae2d77",
"created_at": "2026-08-06 17:52:02.436651+00:00",
"db_md5": "b327011e59b5752201420cf504535a71",
"submitter": "TZZheng",
"github_login": "TZZheng",
"run_id": "f9d8db56-fa8c-444d-95c9-1ae7060dafe9",
"tier": "easy",
"dataset_id": "e1",
"status": "succeeded"... | {
"score": {
"mean_loss": 2.942264437675476,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.07333333045244217
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 3.976949691772461,
"example_count": 100,
... |
06f87a9a-7513-4835-ae87-0314954f3322 | easy | erdavis0 | 2026-08-25 01:44:04.695226+00:00 | succeeded | ae412f48bd9234888fd7bb68dc2dc649ac54245d371d4ab6f9d7476293bb40ba | 24,120 | null | """Matched categorical-boundary panel: C2 straight-through 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": "06f87a9a-7513-4835-ae87-0314954f3322",
"created_at": "2026-08-25 01:44:04.695226+00:00",
"db_md5": "ea525cb9db225168d76b8b0dc9c9bc4a",
"submitter": "Ethan",
"github_login": "erdavis0",
"run_id": "3094634f-728a-4f4c-b749-4db6e26d7724",
"tier": "easy",
"dataset_id": "e4",
"status": "succeeded",... | {
"score": {
"mean_loss": 2.074800540117746,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.006875000012417635
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.0758531874664756,
"example_count": 120... |
06fb7ed5-a358-4533-bf4f-b81ba1bf4d4d | easy | DDanlov | 2026-08-26 01:00:24.201501+00:00 | succeeded | 884d857bb54841f090892f3d83235e3476deeb3beeb522006c54542025100b39 | 34,315 | null | import math
import time
from typing import Optional, Tuple, Any, Dict, List, Union
import torch
import torch.nn as nn
import torch.nn.functional as F
from torch.optim import Optimizer
try:
from benchmark.api import ModelSpec, OptimizerBundle, OptimizerSpec, Submission, assert_model_state
except ImportError:
tr... | {
"id": "06fb7ed5-a358-4533-bf4f-b81ba1bf4d4d",
"created_at": "2026-08-26 01:00:24.201501+00:00",
"db_md5": "95b53fa55c39e6e97751684a36b663dd",
"submitter": "DDanlov",
"github_login": "DDanlov",
"run_id": "2af2cd95-6b93-4a4b-be08-db3b815963f0",
"tier": "easy",
"dataset_id": "e5",
"status": "succeeded"... | {
"score": {
"mean_loss": 2.289891140163818,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0141666666790843
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.2918686952291583,
"example_count": 600,
... |
07051861-06a7-409a-8a9e-6246229c61a6 | easy | erdavis0 | 2026-08-29 01:34:09.805526+00:00 | succeeded | eadd19360be6ecdb82c0d25f4d3447d162d2edd2acd789ba5e873fafaac8fc3b | 19,602 | null | """Four-code N-conditioned circulant operator over an X-initialized tape.
Public delimiters only route opaque categories into independent right-aligned
N, X, and T tapes. N makes a straight-through hard choice among four globally
learned cyclic operators; X alone initializes recurrent state; T is consulted
only after... | {
"id": "07051861-06a7-409a-8a9e-6246229c61a6",
"created_at": "2026-08-29 01:34:09.805526+00:00",
"db_md5": "be2bf3989baac5e04978d28d9e6dbbb4",
"submitter": "Ethan",
"github_login": "erdavis0",
"run_id": "0a93fcef-26b7-4289-901e-ce0bd7fadccc",
"tier": "easy",
"dataset_id": "e7",
"status": "succeeded",... | {
"score": {
"mean_loss": 6.587722539901733,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.12321746721863747
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 8.164097785949707,
"example_count": 85,
... |
070bb68b-9311-47d3-89e6-d1e36973fa6d | easy | richardcepka | 2026-08-13 22:22:56.423617+00:00 | succeeded | b846a7250df6e324ee11027ff09c3eeaea89baa914a013fa1277935e917c6096 | 43,381 | null | """Recurrent digit-register model for One Layer Deeper.
A looped transformer whose recurrent state is a decimal digit register that is
re-quantised through a ten-entry codebook on every iteration. Nothing inside
the loop depends on the iteration index and no parameter is indexed by an
absolute slot position, so the e... | {
"id": "070bb68b-9311-47d3-89e6-d1e36973fa6d",
"created_at": "2026-08-13 22:22:56.423617+00:00",
"db_md5": "ccf53c7b791dbfd728b518e099162c75",
"submitter": "Richard Cepka",
"github_login": "richardcepka",
"run_id": "a3d250b4-0e64-48f8-b154-12bd7a7f6f67",
"tier": "easy",
"dataset_id": "e5",
"status": ... | {
"score": {
"mean_loss": 2.571173965796694,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0008333333535119891
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.560807466506958,
"example_count": 600... |
07199f0a-267c-490a-8b03-5a8ee5131472 | easy | benjaminW2025 | 2026-08-19 05:35:37.728414+00:00 | succeeded | bd739008db9df0cd960bec65df6f11308f15e3897a93d22de3fe4fd8804e4501 | 4,712 | null | """Phase 0 U4: four untied full-token Transformer 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 = 256
NUM_HEADS ... | {
"id": "07199f0a-267c-490a-8b03-5a8ee5131472",
"created_at": "2026-08-19 05:35:37.728414+00:00",
"db_md5": "050481e04268242cf27219a4925ae8e9",
"submitter": "benjawesome",
"github_login": "benjaminW2025",
"run_id": "89caf9db-96bc-4930-a999-699e1972408c",
"tier": "easy",
"dataset_id": "e3",
"status": "... | {
"score": {
"mean_loss": 2.2778620636998,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.012500000149011613
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.4181331806840842,
"example_count": 800,
... |
07220007-e1eb-4785-af17-6182bd40542e | easy | viridale | 2026-08-24 22:32:40.013417+00:00 | succeeded | 0738f208245df41b0eae707674b394a1ea020b765f13fbab797d474369a4816f | 16,107 | null | """Pure pair-table Transformer with endpoint depth constraints.
hypothesis: v863's T1-only objective lets a 2.25M-parameter operator memorize 147
direct rows; using every ordinary supplied endpoint through the same tied state
transition makes interpolation inconsistent while retaining direct T1 pressure
axis: ... | {
"id": "07220007-e1eb-4785-af17-6182bd40542e",
"created_at": "2026-08-24 22:32:40.013417+00:00",
"db_md5": "9408aad3113578966ce0246affc3c070",
"submitter": "priormancer",
"github_login": "viridale",
"run_id": "6a66f3d2-f34c-459e-abba-96cd7921d7b0",
"tier": "easy",
"dataset_id": "e6",
"status": "succe... | {
"score": {
"mean_loss": 2.005194664001465,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.3081081286072731
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 1.6834585666656494,
"example_count": 60,
... |
0729c933-ecde-4883-8b75-cb9d80f1ab11 | easy | sapient-sapiens | 2026-08-25 09:13:17.289546+00:00 | succeeded | 7208ff68edc245a2ed06805eae26d049ca256f64347cfa427d63edee88e257ba | 19,614 | null | """Standalone nested inner/outer recurrence candidate.
Representation is categorical only: ``E_token + E_group + E_leftrel``. No
field is decoded into a number, no field value sets depth, halting, or any
mixing weight, and the only supervised target is the final answer.
Structure
---------
One prompt encoder produce... | {
"id": "0729c933-ecde-4883-8b75-cb9d80f1ab11",
"created_at": "2026-08-25 09:13:17.289546+00:00",
"db_md5": "398087955d34db7496d6658e7d91be3c",
"submitter": "Ertondy",
"github_login": "sapient-sapiens",
"run_id": "9c6b2dc9-9f0b-469a-b8e3-b44cf010934d",
"tier": "easy",
"dataset_id": "e5",
"status": "su... | {
"score": {
"mean_loss": 2.239262359429238,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.007916666691501935
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.2383824314237177,
"example_count": 600... |
072a36df-a0ea-4729-8445-33e3520591b5 | easy | DDanlov | 2026-08-21 21:12:41.512881+00:00 | succeeded | da17f9effeb43b5107aa76eccc87ffa7d95d5547066dfa691f8577feb377bfeb | 27,612 | null | from __future__ import annotations
import math
import time
from typing import Any, Optional, Tuple, Dict
import torch
import torch.nn as nn
import torch.nn.functional as F
from torch.optim import Optimizer
try:
from benchmark.api import ModelSpec, OptimizerBundle, OptimizerSpec, Submission, assert_model_state
exce... | {
"id": "072a36df-a0ea-4729-8445-33e3520591b5",
"created_at": "2026-08-21 21:12:41.512881+00:00",
"db_md5": "fac13cdfcb6a25cc19dea510d2ace40f",
"submitter": "DDanlov",
"github_login": "DDanlov",
"run_id": "7f0f1a82-1a44-497b-9c73-e63857240394",
"tier": "easy",
"dataset_id": "e5",
"status": "succeeded"... | {
"score": {
"mean_loss": 2.8145407672452416,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.004583333333333333
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.823344338635158,
"example_count": 600... |
07300605-c2c3-496f-9bd5-42d2665d32c1 | easy | 0Chris5R | 2026-08-02 16:12:48.213524+00:00 | succeeded | a940aabfacbe52b32e9cd0e1cc3f3477cf8e848ca5c10174289fd51fe00706f2 | 24,899 | null | """Shared-digit Neural GPU with late full-objective EMA evaluation.
The model aligns decimal digits by significance, forms a learned low-rank
second-order feature map of the current state, and evolves a two-dimensional
workspace with the original convolutional-GRU equations. Each arithmetic
transition applies the sam... | {
"id": "07300605-c2c3-496f-9bd5-42d2665d32c1",
"created_at": "2026-08-02 16:12:48.213524+00:00",
"db_md5": "6965304721cd25c894abebf0c462c63f",
"submitter": "Chris ",
"github_login": "0Chris5R",
"run_id": "a5bc9121-0fc5-4207-b25a-3e9fe5adc450",
"tier": "easy",
"dataset_id": "e5",
"status": "succeeded"... | {
"score": {
"mean_loss": 2.684601475227881,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0341666666790843
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.518601326664467,
"example_count": 600,
... |
0731d9e2-a9c4-4a45-ba70-580e6683b04d | easy | lpbb | 2026-08-18 10:01:33.722983+00:00 | succeeded | 4eebbee03109c4c94e2a6c00175021fb8ba1e51e0c4048ca29d432e56619db8f | 4,467 | null | """S0 — control: same shape as S1 but 24 untied layers (fixed depth, no weight tying).
Measures the 'tying tax'; by construction its depth cannot follow T."""
from __future__ import annotations
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
ModelSpec,
Optimi... | {
"id": "0731d9e2-a9c4-4a45-ba70-580e6683b04d",
"created_at": "2026-08-18 10:01:33.722983+00:00",
"db_md5": "25e42d22ba03a93f776aa146fbfbf5f8",
"submitter": "Lpbb",
"github_login": "lpbb",
"run_id": "0974227f-c9a5-443c-be47-c781e06f11a5",
"tier": "easy",
"dataset_id": "e1",
"status": "succeeded",
"s... | {
"score": {
"mean_loss": 3.9619336128234863,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.041666666977107525
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 4.33132791519165,
"example_count": 100,... |
0735ae5f-1358-47cd-a30b-20d9eafdd841 | easy | isaac0804 | 2026-08-06 15:17:23.115476+00:00 | succeeded | 0b586ef0b0124c2dfe172e1615848f1b7ee4636bfc3ad28722be7bd8e2829f8d | 8,357 | null | """E3 candidate #5: discrete vector-quantization bottleneck, a structural
(not soft) information limit on how much per-example-specific state the
model can carry forward.
Everything tried so far this session to discourage the joint-training
memorization collapse was a *soft* discouragement -- dropout, SAM,
extreme wei... | {
"id": "0735ae5f-1358-47cd-a30b-20d9eafdd841",
"created_at": "2026-08-06 15:17:23.115476+00:00",
"db_md5": "1cabffaab4fb733934d5e475bdd22c43",
"submitter": "Isaac Yong",
"github_login": "isaac0804",
"run_id": "24bdd669-b4d9-41a1-aed2-1ba8fdb8a49a",
"tier": "easy",
"dataset_id": "e3",
"status": "succe... | {
"score": {
"mean_loss": 4.125696267429998,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.006250000013969839
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 4.037078032936927,
"example_count": 800,... |
073c6b00-31ee-472f-880b-f89419e62bab | easy | khushidahi | 2026-08-05 22:52:33.999434+00:00 | succeeded | fbd1659041398ddd50514d8fe1fe57566fb2c026ca6b001b666b37d235ca721d | 16,106 | 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": "073c6b00-31ee-472f-880b-f89419e62bab",
"created_at": "2026-08-05 22:52:33.999434+00:00",
"db_md5": "e29183aaa182ca65d59a689fea1203e4",
"submitter": "khushidahi",
"github_login": "khushidahi",
"run_id": "75683027-f892-4337-9018-2dc05f10a416",
"tier": "easy",
"dataset_id": "e5",
"status": "succ... | {
"score": {
"mean_loss": 2.2026573419570923,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.002916666795499623
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.199707508087158,
"example_count": 600... |
0743aee2-6c3f-4067-8060-322ca0b55c87 | easy | sapient-sapiens | 2026-08-07 08:08:45.083938+00:00 | succeeded | 6689b33057c4522b912c24c8302d89f2fa903d1ed61ad52548b8f70d60d79215 | 12,464 | 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": "0743aee2-6c3f-4067-8060-322ca0b55c87",
"created_at": "2026-08-07 08:08:45.083938+00:00",
"db_md5": "a30cbeab0811362a9107bf42db0134aa",
"submitter": "Ertondy",
"github_login": "sapient-sapiens",
"run_id": "0f8ceea6-279f-4a1d-89e0-e4e903f44d6b",
"tier": "easy",
"dataset_id": "e1",
"status": "su... | {
"score": {
"mean_loss": 4.306292591844116,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.05833333214124044
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 4.75090217590332,
"example_count": 100,
... |
07594419-2c5c-4a71-b1fb-e0c461971c4e | easy | DDanlov | 2026-08-26 18:31:12.093695+00:00 | failed | 2a5cc742d3e16d8fbc9f6f081c08e14c2d4c2792ed55d62c3f315878277d3ef8 | 29,496 | null | import math
import time
from typing import Any
import torch
import torch.nn as nn
import torch.nn.functional as F
from torch.optim import Optimizer
try:
from onelayerdeeper.api import Submission, OptimizerBundle, ModelSpec, OptimizerSpec
except ImportError:
Submission, OptimizerBundle, ModelSpec, OptimizerSpec... | {
"id": "07594419-2c5c-4a71-b1fb-e0c461971c4e",
"created_at": "2026-08-26 18:31:12.093695+00:00",
"db_md5": "e134c3368bbdd6780239a0ab087e6f61",
"submitter": "DDanlov",
"github_login": "DDanlov",
"run_id": "fc26b4a2-4511-4602-b921-17b601e66f3b",
"tier": "easy",
"dataset_id": "e5",
"status": "failed",
... | null
|
0760f0fc-5cc6-4ca3-82de-62edb8253acc | easy | shreyash-chonkie | 2026-08-26 18:48:20.852511+00:00 | succeeded | b3ff5313129b98f45be39e18e2ba2c921301c61338af4515a8f86591beea6f7c | 9,852 | null | """Exact consensus between two shared ternary reduction trees."""
from __future__ import annotations
import math
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
assert_model_state,
)
STATE_... | {
"id": "0760f0fc-5cc6-4ca3-82de-62edb8253acc",
"created_at": "2026-08-26 18:48:20.852511+00:00",
"db_md5": "dda4543a961eb3becf38e2f4b6607fe1",
"submitter": "Shreyash",
"github_login": "shreyash-chonkie",
"run_id": "cf669389-4c9a-464a-b5a5-9c7baf1683f9",
"tier": "easy",
"dataset_id": "e1",
"status": "... | {
"score": {
"mean_loss": 2.2332375049591064,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.019999999552965164
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.193472146987915,
"example_count": 100... |
0762c206-a87b-4f10-9d84-363cbc0081d2 | easy | pepinu | 2026-08-21 06:37:56.512055+00:00 | succeeded | 8f2604323b0935ddcc7d49ce89cf4a60dda7735abe7afeae832b0240f7769930 | 40,010 | null | """Squaring mod N as a learned permutation on CRT components.
WHY THIS SHAPE. (Z_N, x) is a commutative monoid, so every finite-dimensional
linear representation splits into multiplicative characters and squaring is
diagonal there -- the component basis is the only one in which this operation is
shallow. Two consequen... | {
"id": "0762c206-a87b-4f10-9d84-363cbc0081d2",
"created_at": "2026-08-21 06:37:56.512055+00:00",
"db_md5": "67bdec213264108b780b80da455f8b04",
"submitter": "pepinu",
"github_login": "pepinu",
"run_id": "ccc10b9c-9323-48be-a703-36a2373d1e4a",
"tier": "easy",
"dataset_id": "e2",
"status": "succeeded",
... | {
"score": {
"mean_loss": 0.014270015060901642,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 1.0
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 0.014291377738118172,
"example_count": 300,
... |
0762e8b4-8cb2-4e27-8a71-de6768c45599 | easy | DDanlov | 2026-08-31 21:31:27.691676+00:00 | succeeded | a834b473d2c7a66277a9681de0ab84dc6543109a9f9ca71975f9b623fb04fa83 | 32,410 | null | """
Unified Continuous Hopfield Equilibrium Propagation Deliberation Engine
=======================================================================
Architecture:
- Deliberation: Continuous Hopfield Multi-Layer Energy System (L=4, D=768, d_ffn=2304, H=16)
- Integration: In-Place Strang-Split Conformal Symplectic Velocit... | {
"id": "0762e8b4-8cb2-4e27-8a71-de6768c45599",
"created_at": "2026-08-31 21:31:27.691676+00:00",
"db_md5": "4a6b4a6d70ce2b6b2acf024aad858847",
"submitter": "DDanlov",
"github_login": "DDanlov",
"run_id": "5a55a4d2-588b-41ff-a53a-02cebe692c78",
"tier": "easy",
"dataset_id": "e4",
"status": "succeeded"... | {
"score": {
"mean_loss": 3.1315937132574168,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0016666666666666668
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 3.149896888098553,
"example_count": 12... |
07745ca4-47f9-4441-bbcd-b0f1fdf1bf8c | easy | sapient-sapiens | 2026-08-15 08:12:45.654789+00:00 | succeeded | 28b8c3ad9f92779ffc1a2d2038d8bfff5d6d8978e20f4b7d2e607f79a8526f9a | 14,711 | null | """Narrow/deep looped Transformer with optional latent workspace tokens.
Research candidate for sampled-modulus fixed-T=2. The encoder builds immutable
prompt memory. A stack of learned operator blocks updates a mutable latent
stream, and the whole stack is reused for each task recurrence. Optional blank
workspace ... | {
"id": "07745ca4-47f9-4441-bbcd-b0f1fdf1bf8c",
"created_at": "2026-08-15 08:12:45.654789+00:00",
"db_md5": "2c9e1b46e1498fc84fe972355664d789",
"submitter": "Ertondy",
"github_login": "sapient-sapiens",
"run_id": "64e1b358-07ca-4305-9a19-a897d0bd957a",
"tier": "easy",
"dataset_id": "e3",
"status": "su... | {
"score": {
"mean_loss": 2.1249103851504243,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.01375
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.124263847882275,
"example_count": 800,
"... |
077ad7bc-b93f-47a6-99c6-e1f214a51152 | easy | DDanlov | 2026-08-20 22:41:25.304101+00:00 | failed | 100b533dd539dfd232463d01472bd7a3e37cadfa8def6b772e22e621846f1e81 | 21,250 | null | from __future__ import annotations
import math
import time
from typing import Any, Optional, Tuple
import torch
import torch.nn as nn
import torch.nn.functional as F
from torch.optim import Optimizer
try:
from benchmark.api import ModelSpec, OptimizerBundle, OptimizerSpec, Submission, assert_model_state
except Imp... | {
"id": "077ad7bc-b93f-47a6-99c6-e1f214a51152",
"created_at": "2026-08-20 22:41:25.304101+00:00",
"db_md5": "999f031f36a0d9c1e53bde8f3289cc0b",
"submitter": "DDanlov",
"github_login": "DDanlov",
"run_id": "2c2803a2-8827-4ac3-a3c6-22a7c7410352",
"tier": "easy",
"dataset_id": "e5",
"status": "failed",
... | null
|
077f9a33-790b-4873-9efc-0d14910c7121 | easy | bk3015 | 2026-08-02 23:07:09.264424+00:00 | succeeded | 7c0cb0724b384c3611ab0b05d306525b92f99b8eb459b96e16f3dd9bc15d4f5b | 9,756 | null | """Standalone One Layer Deeper submission for Arithmetic Lab's recurrent model.
This file intentionally depends only on PyTorch and the competition's public
``benchmark`` API. Upload this file by itself as ``submission.py``.
"""
from __future__ import annotations
import math
import torch
import torch.nn.functional... | {
"id": "077f9a33-790b-4873-9efc-0d14910c7121",
"created_at": "2026-08-02 23:07:09.264424+00:00",
"db_md5": "898a67a5fa4eb070febd2e73bc50e1c6",
"submitter": "bk3015",
"github_login": "bk3015",
"run_id": "98cd147e-585f-41fc-baf8-82010865b932",
"tier": "easy",
"dataset_id": "e1",
"status": "succeeded",
... | {
"score": {
"mean_loss": 2.791366457939148,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0283333333209157
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 3.5364553928375244,
"example_count": 100,
... |
0792c1e6-7a68-4f5c-9b5c-730628acd025 | easy | erdavis0 | 2026-08-27 02:26:29.683140+00:00 | succeeded | f0f048421bab235b2c1c7f2958a7cff30fbcaa7cef983b31931d270fd5484080 | 31,301 | null | """Typed-prefix reversible transition B2: canonical sequence mixture.
Public delimiter syntax routes opaque N, X, and T fields into three strict
functional roles. X initializes only an evolving auxiliary tape, N initializes
only an invariant environment lane, and T reaches only a learned monotone
prefix controller. ... | {
"id": "0792c1e6-7a68-4f5c-9b5c-730628acd025",
"created_at": "2026-08-27 02:26:29.683140+00:00",
"db_md5": "84a6cfec2208ae11b2788eb24985c9f5",
"submitter": "Ethan",
"github_login": "erdavis0",
"run_id": "ba421873-ce6f-42fe-b297-4a77662dc33a",
"tier": "easy",
"dataset_id": "e1",
"status": "succeeded",... | {
"score": {
"mean_loss": 7.602046489715576,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.08999999612569809
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 9.805127143859863,
"example_count": 100,
... |
0795f905-177b-45aa-8807-7f992499cf5e | easy | jordanrubin | 2026-08-03 20:45:05.550685+00:00 | succeeded | 09586f90ffa296e3bdbb25e6a0b059c7579344900d2f7d9eb12b0e902c4848f6 | 15,390 | null | """Bidirectional transformer, D=256 x 3 blocks, grokking-tuned AdamW.
Design notes:
- Inputs are padded to config.max_seq_len inside forward and logits sliced
back, so every training batch presents one static shape (compile/cudagraph
friendly, and uniform kernel shapes even in eager).
- Attention mask gets an iden... | {
"id": "0795f905-177b-45aa-8807-7f992499cf5e",
"created_at": "2026-08-03 20:45:05.550685+00:00",
"db_md5": "ef7d67d872a65aa96e81ca53991ab27d",
"submitter": "Jordan Rubin",
"github_login": "jordanrubin",
"run_id": "c5c57141-799e-4ea8-8a9e-e214fc33cee3",
"tier": "easy",
"dataset_id": "e2",
"status": "s... | {
"score": {
"mean_loss": 3.1677775590470576,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.26562499192853767
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 5.237987944196464,
"example_count": 300,... |
07a17433-0b80-45b0-8eb3-bcbbdcadbbbe | easy | ehonig | 2026-08-20 08:29:53.195131+00:00 | succeeded | a4048801a3630288af848179d22c24b711f63ba46c25598803c9b426845df353 | 23,246 | 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": "07a17433-0b80-45b0-8eb3-bcbbdcadbbbe",
"created_at": "2026-08-20 08:29:53.195131+00:00",
"db_md5": "7ad763d9aaa7d1ef866289ec0b535bb6",
"submitter": "Edouardo Honig",
"github_login": "ehonig",
"run_id": "6afb35ac-02ac-46f8-965f-69da3f9aa07b",
"tier": "easy",
"dataset_id": "e1",
"status": "succ... | {
"score": {
"mean_loss": 1.9215565919876099,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.05166666582226753
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 1.8625059127807617,
"example_count": 100... |
07a7ed85-e3b1-41b0-8ab9-ff92c9f4101d | easy | DDanlov | 2026-08-14 08:03:42.830442+00:00 | succeeded | 07bdd808d3da7c87f8f41193fb0b7eadc4e210b41784a54cc5ed67edeaad0ed7 | 21,977 | 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": "07a7ed85-e3b1-41b0-8ab9-ff92c9f4101d",
"created_at": "2026-08-14 08:03:42.830442+00:00",
"db_md5": "cbd45f88286e55da918950a1685d7990",
"submitter": "DDanlov",
"github_login": "DDanlov",
"run_id": "48fdee60-8c77-4acf-8145-129fc39c94f6",
"tier": "easy",
"dataset_id": "e3",
"status": "succeeded"... | {
"score": {
"mean_loss": 2.8293695195193873,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.018125000000000002
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.82938854805702,
"example_count": 800,... |
07aeaa60-f359-4437-b664-79cd4ab68033 | easy | jordanrubin | 2026-08-11 19:11:04.035101+00:00 | succeeded | 0f93ed4d07f351beca91ef2e290a0c44323a16a8090f3dc0d883cf7155952d62 | 21,702 | 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": "07aeaa60-f359-4437-b664-79cd4ab68033",
"created_at": "2026-08-11 19:11:04.035101+00:00",
"db_md5": "318216f9e8e3f7335c8009581f1e16cc",
"submitter": "Jordan Rubin",
"github_login": "jordanrubin",
"run_id": "894a1ade-f34b-4225-a32f-ed682f4f3475",
"tier": "easy",
"dataset_id": "e5",
"status": "s... | {
"score": {
"mean_loss": 2.5780992363668176,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.007916666489715377
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.1558708974866057,
"example_count": 60... |
07b0445f-135c-4ac7-9f5c-f3840f299a5f | easy | KaustubhKumar05 | 2026-08-26 13:09:04.396173+00:00 | succeeded | 2885fe16055b9736c8555fd2ee63bda5851a77308105d9e7379b5056884459e9 | 12,762 | null | """Neural-GPU-style conv GRU, v2s: v2 + speedrun pack.
- LR: warmup, hold max to 85% of wall-clock, steep late cliff (2026 schedule theory)
- identity-at-init: cell output projection zero-initialized
- batch reuse x6 (multi-pass theory)
Base (v2):
Changes over v1 (each backed by measured results in the literature):
... | {
"id": "07b0445f-135c-4ac7-9f5c-f3840f299a5f",
"created_at": "2026-08-26 13:09:04.396173+00:00",
"db_md5": "a4af586418bd0d3e81ad6142da57cace",
"submitter": "koz",
"github_login": "KaustubhKumar05",
"run_id": "ed023c8f-4adb-47b3-a6d9-db7496b3836d",
"tier": "easy",
"dataset_id": "e3",
"status": "succee... | {
"score": {
"mean_loss": 8.044128408824594,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.00875
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 8.06164214287077,
"example_count": 800,
"ex... |
07b2873a-c3e0-4832-b313-d4f19711a887 | easy | alirezashirvani-jr | 2026-08-31 07:33:15.188376+00:00 | succeeded | c252ce15d535434ae9695254770a32b96e97a7120717f3216a16792d4ff23118 | 34,903 | null | from __future__ import annotations
import math
import time
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
TokenLossBatch,
assert_model_state,
)
TRACK_FLOOR = 3
TRACK_CEILING = 96
STAGE_B... | {
"id": "07b2873a-c3e0-4832-b313-d4f19711a887",
"created_at": "2026-08-31 07:33:15.188376+00:00",
"db_md5": "b9699e2f9827d45ad6df7d21eed39459",
"submitter": "alirezashirvani-jr",
"github_login": "alirezashirvani-jr",
"run_id": "33db49cc-3819-46da-9ee6-dd3777d456f5",
"tier": "easy",
"dataset_id": "e8",
... | {
"score": {
"mean_loss": 0.001376092026475817,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 1.0
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 0.001336814952082932,
"example_count": 85,
"... |
07b2b648-7232-40bd-b033-40d5c18cd8b0 | easy | erdavis0 | 2026-08-22 07:21:31.135415+00:00 | succeeded | 4052302344c2ee5544752d8dd2b682cc62f86ed61a7d43e055b398661db547e4 | 4,774 | null | """Masked butterfly-style global mixing over sequence positions."""
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
HIDDEN... | {
"id": "07b2b648-7232-40bd-b033-40d5c18cd8b0",
"created_at": "2026-08-22 07:21:31.135415+00:00",
"db_md5": "7b255a5697e39c756d9c195ef0c4c794",
"submitter": "Ethan",
"github_login": "erdavis0",
"run_id": "0c538c2e-bcbf-414b-b7b8-dfb01f7e3f75",
"tier": "easy",
"dataset_id": "e5",
"status": "succeeded",... | {
"score": {
"mean_loss": 9.78483772277832,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.01000000024214387
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 9.842329978942871,
"example_count": 600,
... |
07b3c831-b123-4389-9627-4672233a953f | easy | jayyala | 2026-08-16 23:02:51.315162+00:00 | succeeded | 9fb85c86563736265af0d6681567d86fe953dabca7a4c104c60db37f4450ec4e | 14,657 | null | """A scalable recurrent machine with canonical categorical byte states.
Each integer is represented by a short little-endian tape of 256-way slots.
One tied bidirectional mixing cell learns the unknown state transition from the
current tape and the prompt context, then its categorical output is fed back
for exactly T ... | {
"id": "07b3c831-b123-4389-9627-4672233a953f",
"created_at": "2026-08-16 23:02:51.315162+00:00",
"db_md5": "abe23542b6010fc06d199e62298e64a8",
"submitter": "jayyala",
"github_login": "jayyala",
"run_id": "5007a6ff-381b-4cbb-9a4e-ebb495dd36f8",
"tier": "easy",
"dataset_id": "e2",
"status": "succeeded"... | {
"score": {
"mean_loss": 8.595782279968262,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.7602083683013916
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 4.924506187438965,
"example_count": 300,
... |
07bd8b86-fdac-44d9-8fb3-d9ec32cb33ce | easy | shirvani-jr | 2026-08-22 13:01:17.404259+00:00 | succeeded | 1398bd08847c5eaab4959d387d99f9b6d5d587478909c52a6d433aa99a043c89 | 23,140 | null | """Looped exact-arithmetic digit machine.
The recurrent state is the decimal digit string of the current residue. One
loop application performs one modular squaring, so depth extrapolates in T by
construction. Every functional primitive — digit-by-digit product, full
adder, full subtractor, digit comparator, and com... | {
"id": "07bd8b86-fdac-44d9-8fb3-d9ec32cb33ce",
"created_at": "2026-08-22 13:01:17.404259+00:00",
"db_md5": "a1310d423bb5360fa56d444443de24fa",
"submitter": "Ali",
"github_login": "shirvani-jr",
"run_id": "c5bac41b-b08f-4c33-878c-acaeb5a51b23",
"tier": "easy",
"dataset_id": "e1",
"status": "succeeded"... | {
"score": {
"mean_loss": 0.03318208456039429,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 1.0
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 0.033158864825963974,
"example_count": 100,
"... |
07c4550c-4445-413b-89c4-bb1340cc590f | easy | AFeinfield | 2026-08-18 23:38:21.131362+00:00 | succeeded | dd6fa161e518fbc52f4ee06609ebf626a2e6547dc0d13279f3c9a2df6e87d74f | 13,788 | null | """083: conv kernel 5 -> 9 in the multiply sublayer.
Direct test of the mechanism story: the conv gathers neighbouring digits
so the bilinear form can multiply them, so a wider kernel should let more
digit PAIRS interact per iteration.
Original 073 notes follow.
073: BILINEAR multiply + K=5 agreement heads.
The multi... | {
"id": "07c4550c-4445-413b-89c4-bb1340cc590f",
"created_at": "2026-08-18 23:38:21.131362+00:00",
"db_md5": "7f429cd8b404c5ed7219db333f350708",
"submitter": "AFeinfield",
"github_login": "AFeinfield",
"run_id": "2acd0703-69d7-49e6-8745-a87ad7e3f26e",
"tier": "easy",
"dataset_id": "e2",
"status": "succ... | {
"score": {
"mean_loss": 13.677440643310547,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.014375000260770321
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 13.725117683410645,
"example_count": 30... |
07ce2668-754d-4d09-98f1-cdb8fbbd50ef | easy | yashkant | 2026-08-14 16:05:20.004883+00:00 | succeeded | fc0d4127cc1e6c27eb9de3d87cf06bcf62cf6944d81698eca4237617a385f57a | 34,866 | null | """Round1702 three-projections-per-T Dykstra recursion on the D416 carrier.
Each parsed public-T cycle applies an A-B-A sequence through one shared learned
smooth projection, reinjecting the immutable anchor on every call and updating
the corresponding Dykstra correction memory after each projection. All active
cycle... | {
"id": "07ce2668-754d-4d09-98f1-cdb8fbbd50ef",
"created_at": "2026-08-14 16:05:20.004883+00:00",
"db_md5": "abd47fe6480aca014321b617d2495f16",
"submitter": "Yash Kant",
"github_login": "yashkant",
"run_id": "1333f1bf-cb9b-4bd1-86ba-811027e6ad12",
"tier": "easy",
"dataset_id": "e3",
"status": "succeed... | {
"score": {
"mean_loss": 2.122589625194448,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.01937500013038516
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.1261208626117636,
"example_count": 800,... |
07e0bf59-ccc9-4c58-b914-b8b16b3bb590 | easy | AFeinfield | 2026-08-13 00:24:48.211172+00:00 | succeeded | 68c7220cad92b5e8539601d478138acdfd632a4937dd95a78aca49924f62eb6f | 7,733 | null | """Looped (weight-tied) bidirectional Transformer, d=256, 8 iterations.
017: 013 + end-aligned position embedding (answers are tail-aligned; give
each token its distance-from-end so answer slots have stable roles)
to amortize the evaluator's ~200ms/iteration overhead into 8x more steps.
prior. Same parameter count ord... | {
"id": "07e0bf59-ccc9-4c58-b914-b8b16b3bb590",
"created_at": "2026-08-13 00:24:48.211172+00:00",
"db_md5": "5848321a46ad1b0f5c0b6fe4d8822026",
"submitter": "AFeinfield",
"github_login": "AFeinfield",
"run_id": "17410262-ff04-45aa-84a9-691de4d50934",
"tier": "easy",
"dataset_id": "e1",
"status": "succ... | {
"score": {
"mean_loss": 8.89289903640747,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.029999999329447746
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 8.460805892944336,
"example_count": 100,
... |
07e81964-6efc-431e-95d9-355d2a3a8891 | easy | gauravmishra | 2026-08-09 22:52:26.231772+00:00 | succeeded | ff53ecb548c348c010171474eac8b36fa1fa6f95d982ef1c93bd42644de95f1f | 6,079 | null | """Compound-scaled Transformer with distinct blocks and 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_st... | {
"id": "07e81964-6efc-431e-95d9-355d2a3a8891",
"created_at": "2026-08-09 22:52:26.231772+00:00",
"db_md5": "267f9774007ecc00f0abc489cb3ac9b1",
"submitter": "Gaurav Mishra",
"github_login": "gauravmishra",
"run_id": "b7314715-3c64-45db-b487-4ef71b9ee198",
"tier": "easy",
"dataset_id": "e5",
"status": ... | {
"score": {
"mean_loss": 2.296108735904753,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.005416666691501935
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.313701900131499,
"example_count": 600,... |
07e9d8be-99a6-4839-a70e-67c463e05836 | easy | viridale | 2026-08-03 09:51:48.930993+00:00 | succeeded | a6f367a1a6e212fe65b8cb225585df14fd25251fd4bf920fa8b3b7875d003e64 | 135,536 | null | """Signed phase and second-order temporal-state completion.
hypothesis: v134 signs scalar/source recurrences but still excludes natural
signed coefficients in its existing phase and second-order state tails.
axis: arch
target: revised h1 Max T at least1 while preserving all practice 64/64 gates.
expected_delta: ex... | {
"id": "07e9d8be-99a6-4839-a70e-67c463e05836",
"created_at": "2026-08-03 09:51:48.930993+00:00",
"db_md5": "d9d5155cc4c136a0a43e6b18d445e87f",
"submitter": "priormancer",
"github_login": "viridale",
"run_id": "31b56817-db58-4037-8e8a-e4fa489462ce",
"tier": "easy",
"dataset_id": "e5",
"status": "succe... | {
"score": {
"mean_loss": 0.003015164290974301,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 1.0
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 0.00301516398206988,
"example_count": 600,
"... |
07f425bc-ffab-4fc3-b4e9-9a733a2b2094 | easy | shirvani-jr | 2026-08-28 00:48:27.372683+00:00 | succeeded | 396fee419635addaf93a12d3a1446ceda5cdae5e6fca6dbe436e0fa97b7929b8 | 48,549 | 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": "07f425bc-ffab-4fc3-b4e9-9a733a2b2094",
"created_at": "2026-08-28 00:48:27.372683+00:00",
"db_md5": "6eecdcb811fc7c9142f7f4a70d9f9fa1",
"submitter": "Ali",
"github_login": "shirvani-jr",
"run_id": "9c24cd77-01d1-452d-894a-29dbad276336",
"tier": "easy",
"dataset_id": "e2",
"status": "succeeded"... | {
"score": {
"mean_loss": 2.742504835128784,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.3437500111758709
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 3.7835583686828613,
"example_count": 300,
... |
08004367-2e64-4b76-b9c7-2ab0e82010b2 | easy | ehonig | 2026-08-19 17:08:02.055931+00:00 | succeeded | d634d324bdc5e29a7934f14a81e61192db72f3b1c0330d05c5842b76fa84cc7b | 18,910 | 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": "08004367-2e64-4b76-b9c7-2ab0e82010b2",
"created_at": "2026-08-19 17:08:02.055931+00:00",
"db_md5": "2a7b95e96e9d81f7205f5e61520e936b",
"submitter": "Edouardo Honig",
"github_login": "ehonig",
"run_id": "db0fdf01-0416-4360-966e-947794fd00d1",
"tier": "easy",
"dataset_id": "e3",
"status": "succ... | {
"score": {
"mean_loss": 2.2037601470947266,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.00812499993480742
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.202545404434204,
"example_count": 800,... |
08035cc6-025e-4aa9-944a-8d2dadd0a224 | easy | khushidahi | 2026-08-03 22:43:59.898444+00:00 | succeeded | 46c341540806296a4bfd02fa09fdfa8ae97185e04b15d6c49e23e0c11577052d | 15,179 | 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": "08035cc6-025e-4aa9-944a-8d2dadd0a224",
"created_at": "2026-08-03 22:43:59.898444+00:00",
"db_md5": "9d6b30e803a998e488f9602055897d2f",
"submitter": "khushidahi",
"github_login": "khushidahi",
"run_id": "6b01d888-afc0-4adf-a2b1-a8cd8dce3f2f",
"tier": "easy",
"dataset_id": "e5",
"status": "succ... | {
"score": {
"mean_loss": 2.2093095779418945,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.009583333507180214
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.2078356742858887,
"example_count": 60... |
080c0d9d-8270-4566-ada0-378845d4f3d7 | easy | chad-atexpedient | 2026-08-04 20:17:28.564385+00:00 | succeeded | 6f32bb454a766088b409fd3d165e1a123f256a9224c876f791837f1116421798 | 3,907 | null | """R001: sub-1M five-block bidirectional Glint-style encoder."""
from __future__ import annotations
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
assert_model_state,
)
WIDTH = 128
HEADS = ... | {
"id": "080c0d9d-8270-4566-ada0-378845d4f3d7",
"created_at": "2026-08-04 20:17:28.564385+00:00",
"db_md5": "2e631cc9c06ae8ae4e998074619570da",
"submitter": "chad-atexpedient",
"github_login": "chad-atexpedient",
"run_id": "b4462dde-ca8c-473e-816f-5c271be32d69",
"tier": "easy",
"dataset_id": "e5",
"st... | {
"score": {
"mean_loss": 9.222842041172342,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0083333333581686
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 9.067545471704594,
"example_count": 600,
... |
0818618e-9b5b-4fa2-af49-6f494b0c7a86 | easy | MarcoF1 | 2026-08-13 06:25:22.708584+00:00 | succeeded | 227203c2713008b7f5483c33cc02ff5237f2ad4eaff8ac39fb4a6b2eff3412e2 | 5,871 | null | """Weight-tied looped transformer; latent depth scales with the prompt's T.
Hypothesis treatment arm: one tied block learns "one modular-squaring step" and
is re-applied 2T+2 times, where T is parsed from the prompt's digit tokens.
Input re-injection and per-iteration RMSNorm keep the state bounded so the
recurrence c... | {
"id": "0818618e-9b5b-4fa2-af49-6f494b0c7a86",
"created_at": "2026-08-13 06:25:22.708584+00:00",
"db_md5": "4b81362a922cf91e70e6fe54627dc4a9",
"submitter": "Marco Fleming",
"github_login": "MarcoF1",
"run_id": "f7dc73de-8b17-4e25-9336-cee49aedc194",
"tier": "easy",
"dataset_id": "e1",
"status": "succ... | {
"score": {
"mean_loss": 12.990639209747314,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.009999999776482582
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 17.01241683959961,
"example_count": 100... |
081c1a25-de96-444f-bdb2-860678d2fc07 | easy | KaustubhKumar05 | 2026-08-27 12:09:38.939397+00:00 | succeeded | 0e9f25a4cc7592f3adc331bfdb9a2f3566e06eb44239586933b609e7350a1085 | 14,701 | null | """abacus_sine — SINUSOIDAL activation instead of GELU.
§72: TWENTY candidates in this project and every single one used GELU -- the
activation was inherited from TransformerEncoderLayer and never once varied.
Salah & Yevick 2411.05353 report that grokking on modular arithmetic is
CONTROLLED by the activation function... | {
"id": "081c1a25-de96-444f-bdb2-860678d2fc07",
"created_at": "2026-08-27 12:09:38.939397+00:00",
"db_md5": "445fe6eb1b47727ceb55b42fc155d11d",
"submitter": "koz",
"github_login": "KaustubhKumar05",
"run_id": "054968c5-17c0-4061-97c7-1541097fafd4",
"tier": "easy",
"dataset_id": "e3",
"status": "succee... | {
"score": {
"mean_loss": 3.3006734626209546,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.006875
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 3.2273350566387378,
"example_count": 800,
... |
081c5099-fd41-4e47-a73e-f165cb5ed887 | easy | chad-atexpedient | 2026-08-09 21:56:36.638639+00:00 | succeeded | 786746be3f2ae2ef48aa2f4984d6e97c26e78517477b67c009c60a94cca31e88 | 10,011 | null | """BL7B: step-400 LR 5e-5."""
from __future__ import annotations
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
assert_model_state,
)
LANES = 6
IMMUTABLE_LANES = 3
MUTABL... | {
"id": "081c5099-fd41-4e47-a73e-f165cb5ed887",
"created_at": "2026-08-09 21:56:36.638639+00:00",
"db_md5": "3d2c1c172b15ac9f6c49ffb7fae1ef68",
"submitter": "chad-atexpedient",
"github_login": "chad-atexpedient",
"run_id": "788a9613-4483-48a7-bd23-819889fb4e29",
"tier": "easy",
"dataset_id": "e4",
"st... | {
"score": {
"mean_loss": 2.3514292723531587,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0027083333333333334
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.359180814066394,
"example_count": 12... |
081e85ec-ace9-4375-8ef7-0b3ca1fa78a5 | easy | DDanlov | 2026-08-14 06:52:38.198554+00:00 | succeeded | b2f053da815cca6caf38c882a6905282265bd04ab274e8f5994dd8da7b8af189 | 21,978 | null | import math
from typing import Optional
import torch
import torch.nn as nn
import torch.nn.functional as F
from torch.optim import Optimizer
try:
from benchmark import Submission, assert_model_state, OptimizerBundle
except ImportError:
try:
from client import Submission, assert_model_state, OptimizerBu... | {
"id": "081e85ec-ace9-4375-8ef7-0b3ca1fa78a5",
"created_at": "2026-08-14 06:52:38.198554+00:00",
"db_md5": "09ca3af0c44de8c8a76f45b9915e5889",
"submitter": "DDanlov",
"github_login": "DDanlov",
"run_id": "2b260635-dedf-43d7-aa6d-87cd8b71d13d",
"tier": "easy",
"dataset_id": "e5",
"status": "succeeded"... | {
"score": {
"mean_loss": 2.832217931632606,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0141666666790843
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.8322324517596464,
"example_count": 600,
... |
0820fe75-0270-4c4a-a61b-ebc4490c5b2b | easy | liam-gb | 2026-08-16 19:57:17.481108+00:00 | succeeded | d31a37f219cf4fff7c387213c155ab57fc18b31c931c0bb210a3061269e4d0a4 | 16,563 | null | """rns_pm_gather_ct_widet_l32nore_fb: the grokking-regime probe, arm 1 of 2.
REGIME CHANGES vs rns_pm_gather_ct_widet_l32nore (declared as a bundle: this
is a pre-registered REGIME probe, not an attribution arm): (1) full-batch
training on E1 — batch_size = 600 = the E1 train-set size exactly (drop_last
is true in eve... | {
"id": "0820fe75-0270-4c4a-a61b-ebc4490c5b2b",
"created_at": "2026-08-16 19:57:17.481108+00:00",
"db_md5": "eea8539e8dc6be91674c6b71bf1facb6",
"submitter": "liam-gb",
"github_login": "liam-gb",
"run_id": "5bbbff0f-289c-41c1-b8cb-f2811025be5a",
"tier": "easy",
"dataset_id": "e1",
"status": "succeeded"... | {
"score": {
"mean_loss": 2.3324100971221924,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.15333333611488342
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 3.134223222732544,
"example_count": 100,... |
0823feac-52e2-45d2-97a9-5158776d1ecf | easy | AFeinfield | 2026-08-14 17:58:03.384134+00:00 | succeeded | f3a4c76f54f86b099fd5042d2ccb6f42950b18c55f347f5a832830ad6bfe4d03 | 7,366 | null | """Looped (weight-tied) bidirectional Transformer, d=256, 8 iterations.
039b: e1 cert sprint v2 — d128 + lr 5e-3: memorize by ~30s, grok the rest — 013 skeleton with m1-proven gentle
constants (wd .3, gf .5) + trapezoid LR. Target: 100% on e1's T=1 rung.
to amortize the evaluator's ~200ms/iteration overhead into 8x mo... | {
"id": "0823feac-52e2-45d2-97a9-5158776d1ecf",
"created_at": "2026-08-14 17:58:03.384134+00:00",
"db_md5": "914739aa43fb3cb0090def5ed11fecc0",
"submitter": "AFeinfield",
"github_login": "AFeinfield",
"run_id": "3aef0aef-dd94-4b3d-bfc9-e8e302bd497a",
"tier": "easy",
"dataset_id": "e1",
"status": "succ... | {
"score": {
"mean_loss": 9.330209255218506,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.01833333307877183
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 9.23244571685791,
"example_count": 100,
... |
08347c5f-39b9-409a-9549-5021e5a5b0e9 | easy | lpbb | 2026-08-12 08:27:06.588113+00:00 | succeeded | 48148ea97a1a36b199fb19bc224ef24c7ebf57d0e2d38aa21636aed0c352e1e7 | 24,567 | null | """DIGIT-STATE DISCRETE FLOW (v2) — diffusion-style quantized recurrence, e1.
v1 (submissions/agents/diffusion) put the recurrent state in a learned
codebook. Flaw found on paper: nothing ties step(enc(x)) to enc(f(x)), so
the code space can fragment into per-chain private paths and the coverage
argument (98-99% of e... | {
"id": "08347c5f-39b9-409a-9549-5021e5a5b0e9",
"created_at": "2026-08-12 08:27:06.588113+00:00",
"db_md5": "514122284aa1150d8fec9e21ccebf641",
"submitter": "Lpbb",
"github_login": "lpbb",
"run_id": "17113f68-fc3e-4d0c-a1ca-a8f950197268",
"tier": "easy",
"dataset_id": "e1",
"status": "succeeded",
"s... | {
"score": {
"mean_loss": 0.692231547087431,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.8799999952316284
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 0.048262618482112885,
"example_count": 100... |
083665a5-bef7-42d1-8a8a-76d8eedd0eb1 | easy | jordanrubin | 2026-08-14 16:34:27.286509+00:00 | succeeded | 023e1030251d98f51f73d6cb94a632e5b528dc1697c7cf72f9d6b4d01ae27dae | 14,579 | null | """Recalled neural-GPU transition for the variable-modulus Easy task.
The public decimal fields are represented only as right-origin categorical
registers. One position-invariant local gated step is reused at every place
and microstep. During an outer transition, immutable categorical N and the
current categorical s... | {
"id": "083665a5-bef7-42d1-8a8a-76d8eedd0eb1",
"created_at": "2026-08-14 16:34:27.286509+00:00",
"db_md5": "25d73b639e4d41ff61662718d08247de",
"submitter": "Jordan Rubin",
"github_login": "jordanrubin",
"run_id": "d086c257-597d-4d6e-99d0-59f6f6cd287f",
"tier": "easy",
"dataset_id": "e5",
"status": "s... | {
"score": {
"mean_loss": 2.465743300424127,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.006250000012417635
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.3979661413906936,
"example_count": 600... |
083b9050-40ef-41e7-8f05-178261d2d0f3 | easy | yashkant | 2026-08-16 00:22:18.710787+00:00 | succeeded | 1bc8df6a5e907b3a48e004b2a4c244e7b5b6835c95144812f8aba0bf9e5a6399 | 48,086 | null | """Round1805 logarithmic dyadic carry closure pair.
The exact Round1792 AdamW5 carrier is retained. Learned nonlinear carry
transfer curves are composed across right-aligned N/X token positions by a
binary-lifting closure. Parsed public T allocates only ``bit_length(T)``
ordered composition stages, with offsets cycl... | {
"id": "083b9050-40ef-41e7-8f05-178261d2d0f3",
"created_at": "2026-08-16 00:22:18.710787+00:00",
"db_md5": "378d71bd04ef53dc577c8314795f22fa",
"submitter": "Yash Kant",
"github_login": "yashkant",
"run_id": "38ebb193-e1fb-4978-ab44-dc8e0115e55b",
"tier": "easy",
"dataset_id": "e3",
"status": "succeed... | {
"score": {
"mean_loss": 2.128338637364949,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.01375000014901161
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.1253477197060526,
"example_count": 800,... |
083c9e29-7d62-4f06-822a-799fcf54f92f | easy | CodeReclaimers | 2026-08-18 22:30:04.293347+00:00 | succeeded | d12107e96c7eb0c5b87f90836a677a9f097d59b50f7b9d9f61e7b2216153ea29 | 13,459 | null | """T=1-first field-aware functional-root search.
Four cells share one reversible Feistel mechanism and differ only in how the
visible N field is interpreted: ordinary decimal, reversed decimal, inverse
Gray, or bit-reversed. X and Y remain in the visible 20-bit state coordinate.
Each cell also has a small, zero-initi... | {
"id": "083c9e29-7d62-4f06-822a-799fcf54f92f",
"created_at": "2026-08-18 22:30:04.293347+00:00",
"db_md5": "ee2be1eafc03e9d2321fa079c29427d7",
"submitter": "CodeReclaimers",
"github_login": "CodeReclaimers",
"run_id": "34dd0541-73e6-414b-a163-35bc5a81b24d",
"tier": "easy",
"dataset_id": "e3",
"status... | {
"score": {
"mean_loss": 20.976418157363085,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.010625000004656613
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 20.829617779948002,
"example_count": 80... |
08427a5b-77b4-430e-ab17-fe60ea1a1308 | easy | erdavis0 | 2026-08-26 04:30:21.408573+00:00 | succeeded | f323ea27a95b25c17ef240c64fd37d3915d7f3e1d86c503b6dc262e4235c4217 | 30,352 | null | """Latent-role continuous REV C1: selector-oracle distillation.
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 first e... | {
"id": "08427a5b-77b4-430e-ab17-fe60ea1a1308",
"created_at": "2026-08-26 04:30:21.408573+00:00",
"db_md5": "1658fcec88828e061d8640031c475500",
"submitter": "Ethan",
"github_login": "erdavis0",
"run_id": "a574fa42-ff99-4c31-b63e-640e05a889a1",
"tier": "easy",
"dataset_id": "e1",
"status": "succeeded",... | {
"score": {
"mean_loss": 4.3349034786224365,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.054999999701976776
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 4.475901126861572,
"example_count": 100... |
084868ac-7816-448c-9840-12c890ec5252 | easy | khushidahi | 2026-08-08 00:53:59.911035+00:00 | succeeded | 70d0a0e3464b3dedce66d7ea94eb6c298bb59bc363b414d76ef09bfda66d59e6 | 13,208 | null | """R42 Fourier-number representation experiment.
All input-dependent numeric parsing stays inside differentiable GPU tensor
operations. This is a learned-model representation, not a task solver: it does
not compute modular squaring or create intermediate labels.
"""
from __future__ import annotations
import math
impor... | {
"id": "084868ac-7816-448c-9840-12c890ec5252",
"created_at": "2026-08-08 00:53:59.911035+00:00",
"db_md5": "193ff75715fc01d0ed5f58d853c8e324",
"submitter": "khushidahi",
"github_login": "khushidahi",
"run_id": "548dca46-c633-41ea-95e2-2a5b8a8c98da",
"tier": "easy",
"dataset_id": "e5",
"status": "succ... | {
"score": {
"mean_loss": 2.157176305514735,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.009166666759798924
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.160970449447632,
"example_count": 600,... |
084c8fe9-500c-4900-a56f-0f807a7f17e1 | easy | KaustubhKumar05 | 2026-08-25 15:31:19.583937+00:00 | succeeded | cede8f877bf90dc9d38af3879dc3680a73b6693631c3476085830cfc30a955d0 | 16,027 | null | """final_v1: culmination candidate. v8 (N-FiLM) + Muon momentum 0.5 + tail-EMA.
A small MLP embeds the modulus digits (mean-pooled over the N span) and
emits per-channel scale/shift applied to the cell input and a bias on the
update gate, every iteration. Hypernetwork-lite: the executor is modulated
by N rather than h... | {
"id": "084c8fe9-500c-4900-a56f-0f807a7f17e1",
"created_at": "2026-08-25 15:31:19.583937+00:00",
"db_md5": "4c081115900db9bcea126ae905aeee44",
"submitter": "koz",
"github_login": "KaustubhKumar05",
"run_id": "7b099340-7a31-4192-a015-af3adcfc6ba5",
"tier": "easy",
"dataset_id": "e6",
"status": "succee... | {
"score": {
"mean_loss": 5.352868914604187,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.3216216266155243
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 6.95753288269043,
"example_count": 60,
... |
084ce65f-83ec-48df-91f4-7b1b6b17ccde | easy | Dandandan | 2026-08-05 17:16:10.357439+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": "084ce65f-83ec-48df-91f4-7b1b6b17ccde",
"created_at": "2026-08-05 17:16:10.357439+00:00",
"db_md5": "eaac69fe345f82973fcf392c504a40ff",
"submitter": "Daniël Heres",
"github_login": "Dandandan",
"run_id": "6a60d544-4e27-4be9-a945-a9c6e45ece74",
"tier": "easy",
"dataset_id": "e5",
"status": "suc... | {
"score": {
"mean_loss": 2.787245909823257,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0033333333333333335
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.9658327487552114,
"example_count": 60... |
085361ea-341c-4858-bbd3-d55da1d7e607 | easy | repst | 2026-08-24 20:54:40.683742+00:00 | succeeded | 0f498bc7ec9acee3741b6fb990aeb46afd49e2b46500f63540dbec1b9fa756a7 | 22,257 | null | # R635: lazy two-instruction world model -> learned parallel transducer.
#
# PURPOSE AND RULES BOUNDARY
# --------------------------
# This is an intentionally transparent boundary experiment. During training,
# an MDL-regularized two-instruction modular-ring vocabulary supplies a differentiable
# mixture over latent ... | {
"id": "085361ea-341c-4858-bbd3-d55da1d7e607",
"created_at": "2026-08-24 20:54:40.683742+00:00",
"db_md5": "df41b762da8a7c4623d1c49c9f213f3f",
"submitter": "Ryan Epstein",
"github_login": "repst",
"run_id": "d72d9b2e-2642-49f3-b207-1cca888beb46",
"tier": "easy",
"dataset_id": "e5",
"status": "succeed... | {
"score": {
"mean_loss": 0.009073286802687812,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 1.0
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 0.009113912554687482,
"example_count": 600,
... |
08551f55-d0c8-4d46-b177-30d61c1012c9 | easy | lpbb | 2026-08-12 08:43:52.787079+00:00 | succeeded | 0f5431f2eb4874f1fe1ae671550966658c560716095cf1debdc7603ab4464457 | 25,151 | null | """DIGIT-STATE DISCRETE FLOW (v2) — diffusion-style quantized recurrence, e1.
v1 (submissions/agents/diffusion) put the recurrent state in a learned
codebook. Flaw found on paper: nothing ties step(enc(x)) to enc(f(x)), so
the code space can fragment into per-chain private paths and the coverage
argument (98-99% of e... | {
"id": "08551f55-d0c8-4d46-b177-30d61c1012c9",
"created_at": "2026-08-12 08:43:52.787079+00:00",
"db_md5": "698b340ca5abbb57f9c3a67760bd468a",
"submitter": "Lpbb",
"github_login": "lpbb",
"run_id": "e90822d2-1300-4eed-895b-ce578265adea",
"tier": "easy",
"dataset_id": "e3",
"status": "succeeded",
"s... | {
"score": {
"mean_loss": 3.475030126907802,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.014375000013969838
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 3.4692527571795715,
"example_count": 800... |
085b99d9-ed0a-471e-a5a3-ece655aeaacd | easy | chad-atexpedient | 2026-08-09 21:39:34.917157+00:00 | succeeded | b97e39010bcefe6d24f8bd2bb5495e9d5e53c5327f82f965c46c0efd139c88e9 | 10,027 | null | """BL5A: BL1 post-400 LR 1e-4 continuation."""
from __future__ import annotations
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
assert_model_state,
)
LANES = 6
IMMUTABLE_... | {
"id": "085b99d9-ed0a-471e-a5a3-ece655aeaacd",
"created_at": "2026-08-09 21:39:34.917157+00:00",
"db_md5": "2dcbd32afea71272821b5401c12cfb5e",
"submitter": "chad-atexpedient",
"github_login": "chad-atexpedient",
"run_id": "679dacf0-2952-4fe2-9cad-65d16594be6d",
"tier": "easy",
"dataset_id": "e4",
"st... | {
"score": {
"mean_loss": 2.424318815798382,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0043749999999999995
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.432443422765119,
"example_count": 120... |
085d8932-5a6a-4d0c-a671-a54e1fa79470 | easy | yashkant | 2026-08-13 00:05:56.494929+00:00 | succeeded | fcb6e4f4719005cd2391ce4c5fe2e3637adbe575d6494caf8c70cbf0ad8d441c | 35,038 | null | """Round1542 low-capacity periodic-fold answer/latent state machine.
A width-32 answer/latent pair is refined by one shared hidden-16 MLP once per
public-T macro. The candidate advances each field through a differentiable
triangle-wave fold, giving bounded carry-free wraparound; the equal-work
control uses a conventio... | {
"id": "085d8932-5a6a-4d0c-a671-a54e1fa79470",
"created_at": "2026-08-13 00:05:56.494929+00:00",
"db_md5": "e6e641a83d0464513d5fee5a23fb1f78",
"submitter": "Yash Kant",
"github_login": "yashkant",
"run_id": "ee2b44d5-6574-48c9-a430-ddc6aab485f0",
"tier": "easy",
"dataset_id": "e2",
"status": "succeed... | {
"score": {
"mean_loss": 4.539708852767944,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.33750001434236765
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 7.458319664001465,
"example_count": 300,
... |
085fa5fe-2998-4a7b-8259-98a7bb2f1016 | easy | YuMingQian1234 | 2026-08-07 05:15:12.980798+00:00 | failed | a61dcaedd01b7ae585326b70102e444f37c301ced4e45f9e24c360fca309864a | 18,380 | null | """
Full definition of a GPT Language Model, all of it in this single file.
References:
1) the official GPT-2 TensorFlow implementation released by OpenAI:
https://github.com/openai/gpt-2/blob/master/src/model.py
2) huggingface/transformers PyTorch implementation:
https://github.com/huggingface/transformers/blob/main/s... | {
"id": "085fa5fe-2998-4a7b-8259-98a7bb2f1016",
"created_at": "2026-08-07 05:15:12.980798+00:00",
"db_md5": "ad8f063f37b1cf94bc5fa254e624d17d",
"submitter": "Mingqian Yu",
"github_login": "YuMingQian1234",
"run_id": "c85e8861-aa7f-4228-b639-f2d568cc7167",
"tier": "easy",
"dataset_id": "e1",
"status": ... | null
|
08635b4f-cd8e-4206-923a-4ef7ccbc0e5f | easy | khushidahi | 2026-08-05 22:29:16.999837+00:00 | succeeded | ad06b40c13581d881d49b63253959c8932a1293339b35ddf50cae6c98112e0b5 | 25,699 | null | """R16 hard-transfer hybrid experiments for One Layer Deeper.
Motivated by complementary Hard behavior:
- the generic sequence recurrent model retained seen-N first-rung progress;
- the structured decimal-tape model retained OOD-N first-rung progress.
The dual variants co-train both learned representations, supervise... | {
"id": "08635b4f-cd8e-4206-923a-4ef7ccbc0e5f",
"created_at": "2026-08-05 22:29:16.999837+00:00",
"db_md5": "1c822de5b7c016a75af39f3277594281",
"submitter": "khushidahi",
"github_login": "khushidahi",
"run_id": "ae353b1f-7080-46e3-9fc7-471442716614",
"tier": "easy",
"dataset_id": "e5",
"status": "succ... | {
"score": {
"mean_loss": 2.1793466747344077,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.01250000017384688
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.181122064590454,
"example_count": 600,... |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.