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 |
|---|---|---|---|---|---|---|---|---|---|---|
1c069f07-21e6-4307-a595-e172187896be | easy | khushidahi | 2026-08-05 20:30:58.585078+00:00 | succeeded | 609b4973cd6b33a3e8e8261b352d978bf48f9d74793dbb563d84440372fd8cd7 | 16,674 | 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": "1c069f07-21e6-4307-a595-e172187896be",
"created_at": "2026-08-05 20:30:58.585078+00:00",
"db_md5": "edf9cf570646540fdd872d8235dc4a6a",
"submitter": "khushidahi",
"github_login": "khushidahi",
"run_id": "fa7ff3bc-c56a-459c-a498-a05f0470c7fb",
"tier": "easy",
"dataset_id": "e5",
"status": "succ... | {
"score": {
"mean_loss": 2.2132123708724976,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.01208333345130086
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.211557626724243,
"example_count": 600,... |
1c07cc2c-cf07-4da8-a9db-1b03ea24c711 | easy | DDanlov | 2026-08-20 11:43:21.305497+00:00 | succeeded | 122556183432788ae17eac913c150ad0af81a4bda204b9083f4a3dac23f608c2 | 20,912 | null | from __future__ import annotations
import math
import time
from typing import Any, Optional, Tuple
import torch
import torch.nn as nn
import torch.nn.functional as F
from torch.optim import Optimizer
try:
from benchmark.api import ModelSpec, OptimizerBundle, OptimizerSpec, Submission, assert_model_state
except Imp... | {
"id": "1c07cc2c-cf07-4da8-a9db-1b03ea24c711",
"created_at": "2026-08-20 11:43:21.305497+00:00",
"db_md5": "986e61914330c84720c96bd3b63f5aa5",
"submitter": "DDanlov",
"github_login": "DDanlov",
"run_id": "c93139aa-2262-4f9e-97c1-fbd1df112844",
"tier": "easy",
"dataset_id": "e5",
"status": "succeeded"... | {
"score": {
"mean_loss": 2.2919209834718126,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.012083333345750968
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.2909588086765442,
"example_count": 60... |
1c1686ad-77e3-46c5-82fa-f3a802ac2a8f | easy | DDanlov | 2026-08-31 20:45:36.495894+00:00 | succeeded | e926c9ea581c90232706c29639eb7d3b8d154f993a2c43ea08f1a584f3c0e475 | 32,412 | null | """
Unified Continuous Hopfield Equilibrium Propagation Deliberation Engine
=======================================================================
Architecture:
- Deliberation: Continuous Hopfield Multi-Layer Energy System (L=4, D=768, d_ffn=2304, H=16)
- Integration: In-Place Strang-Split Conformal Symplectic Velocit... | {
"id": "1c1686ad-77e3-46c5-82fa-f3a802ac2a8f",
"created_at": "2026-08-31 20:45:36.495894+00:00",
"db_md5": "656e60223ca73dbc7f71881f853ccacd",
"submitter": "DDanlov",
"github_login": "DDanlov",
"run_id": "0f1cd32f-c377-4032-8011-2ddf1597315b",
"tier": "easy",
"dataset_id": "e4",
"status": "succeeded"... | {
"score": {
"mean_loss": 2.2958458702290194,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.004583333333333333
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.3108860632051087,
"example_count": 12... |
1c1df927-2e10-4e9d-9926-f5684c79dac6 | easy | oupadhyay | 2026-08-07 12:09:34.386506+00:00 | succeeded | 6ba82b5ffcbe4493f3d10e5ab3921b8ba3cd47382832b5f297b1b9da58d147be | 9,181 | null | """Compressed axial v10 with zero-initialized boundary-to-pair attention."""
import torch
import torch.nn.functional as F
from torch import nn
from benchmark import OptimizerBundle, Submission, assert_model_state
PAD, N, X, T, ANS, DIGIT = 0, 2, 3, 4, 5, 7
MAX_STEPS, CHANNELS, STATE_ELEMENTS = 64, 64, 63146
class Con... | {
"id": "1c1df927-2e10-4e9d-9926-f5684c79dac6",
"created_at": "2026-08-07 12:09:34.386506+00:00",
"db_md5": "49103cded34690ac50666ff67e4165a3",
"submitter": "Ojasw Upadhyay",
"github_login": "oupadhyay",
"run_id": "1946c33b-104e-4b90-b39f-298cd76add58",
"tier": "easy",
"dataset_id": "e5",
"status": "s... | {
"score": {
"mean_loss": 3.3535118225241187,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.005416666666666667
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 3.383843628280366,
"example_count": 600... |
1c1f3129-e7e1-42e6-8293-48ad17bf2c21 | easy | Dandandan | 2026-08-04 15:13:42.333627+00:00 | succeeded | 0395e20ff9fde8174aee97c17380596762f596ca524bad48d0528a136cbe4d1b | 17,735 | null | """Multi-start learned spectral recurrence for One Layer Deeper.
Every coefficient that determines a prediction is randomly initialized and
updated end to end. Shared candidate transitions are unrolled to the prompt's
training depth; self-calibration and an integer-boundary loss favor a reusable
functional root withou... | {
"id": "1c1f3129-e7e1-42e6-8293-48ad17bf2c21",
"created_at": "2026-08-04 15:13:42.333627+00:00",
"db_md5": "a03e677ee6ab04b630f92391ec92b4e4",
"submitter": "Daniël Heres",
"github_login": "Dandandan",
"run_id": "26f43b03-5df8-421a-bfdb-08b5ca836e63",
"tier": "easy",
"dataset_id": "e1",
"status": "suc... | {
"score": {
"mean_loss": 1.1271511084487429e-05,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 1.0
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 1.1280336366326082e-05,
"example_count": 100,
... |
1c263d81-37d0-41ab-99a7-c179aeef39bf | easy | khushidahi | 2026-08-07 04:02:28.900335+00:00 | succeeded | fa168ec5415fb577c833424b9c6dffbc02d9303ba3dcf791a3a93fb8afaf4e3e | 17,794 | 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": "1c263d81-37d0-41ab-99a7-c179aeef39bf",
"created_at": "2026-08-07 04:02:28.900335+00:00",
"db_md5": "1c6b16722709842c59f392c931d93d49",
"submitter": "khushidahi",
"github_login": "khushidahi",
"run_id": "a5c9db02-9552-46f9-b788-ff5a16f21a8f",
"tier": "easy",
"dataset_id": "e5",
"status": "succ... | {
"score": {
"mean_loss": 2.158530592918396,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.010416666977107525
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.1583540439605713,
"example_count": 600... |
1c2a328b-b31a-4666-8427-0a378b03651b | easy | sapient-sapiens | 2026-08-20 20:43:04.688973+00:00 | succeeded | a02688fc26f74a1fef3e44eb71fcea9b5571b8beb79eb403da5f7614b45ee5be | 38,230 | null | """Hierarchical learned digit-mixing recurrence with an answer-only 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. T... | {
"id": "1c2a328b-b31a-4666-8427-0a378b03651b",
"created_at": "2026-08-20 20:43:04.688973+00:00",
"db_md5": "401f470a6dfe03a40a53b9a18dc151b8",
"submitter": "Ertondy",
"github_login": "sapient-sapiens",
"run_id": "e0ecb564-6a12-4afd-922e-7922aa23d6d4",
"tier": "easy",
"dataset_id": "e5",
"status": "su... | {
"score": {
"mean_loss": 2.179360631349172,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.005416666691501935
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.1910852942231527,
"example_count": 600... |
1c3609a8-3ecf-4523-bc66-e0367d655bef | easy | DDanlov | 2026-08-08 10:57:16.422832+00:00 | failed | b479b7c2868af5787bd4c74b0c95f585b72c03ebdb60394a23df27de00c260d5 | 15,561 | 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": "1c3609a8-3ecf-4523-bc66-e0367d655bef",
"created_at": "2026-08-08 10:57:16.422832+00:00",
"db_md5": "12490fe74af032d0548f750dbd34ee33",
"submitter": "DDanlov",
"github_login": "DDanlov",
"run_id": "7d982688-201d-4a7c-9c77-a2d27dcb4065",
"tier": "easy",
"dataset_id": "e5",
"status": "failed",
... | null
|
1c3c52c5-9a30-442b-9677-2f7a83213b26 | easy | repst | 2026-08-09 21:45:31.396259+00:00 | succeeded | 36dd679dcb477f7aa19beb9d8cd039a7f44fea0c10036f14f82fddcba0cbe31f | 23,177 | null | """Strict forced-T operator-learning diagnostic for the public prompt contract."""
from __future__ import annotations
from typing import NamedTuple
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,... | {
"id": "1c3c52c5-9a30-442b-9677-2f7a83213b26",
"created_at": "2026-08-09 21:45:31.396259+00:00",
"db_md5": "1db907953e7b39fe8d31af5b24eee25b",
"submitter": "Ryan Epstein",
"github_login": "repst",
"run_id": "eb2c2b82-7330-4fff-9d51-f90f0ddf835a",
"tier": "easy",
"dataset_id": "e1",
"status": "succeed... | {
"score": {
"mean_loss": 7.453524351119995,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.07333333231508732
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 8.313920974731445,
"example_count": 100,
... |
1c4886dc-13fa-4b45-903e-ee371a48a00f | easy | yashkant | 2026-08-14 15:38:23.281314+00:00 | succeeded | b9ba5fa82bdd16907f6f9089a4da8a3fd8150d2fea17d6495bf9807d04e44936 | 28,590 | null | """Round1174 width64 MemGLU plus T-scaled token-y/z crossover."""
from __future__ import annotations
import math
import torch
import torch.nn.functional as F
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
TokenLossBatch,
assert_model_state,
)
from torch import ... | {
"id": "1c4886dc-13fa-4b45-903e-ee371a48a00f",
"created_at": "2026-08-14 15:38:23.281314+00:00",
"db_md5": "da415bff414cf5fdd3f7cb254d929e73",
"submitter": "Yash Kant",
"github_login": "yashkant",
"run_id": "9012e71c-ef06-4805-8e42-3f88d6a05c05",
"tier": "easy",
"dataset_id": "e3",
"status": "succeed... | {
"score": {
"mean_loss": 2.126669136759442,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.01562500013038516
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.1246529224293043,
"example_count": 800,... |
1c49bd6b-88e5-4c63-93c9-07bdb0f64ac3 | easy | DDanlov | 2026-08-12 22:40:07.696334+00:00 | failed | da74ebe078117f7172f81293f5b4a218595df7006a955243fa9374a2bdc913fb | 27,361 | null | from __future__ import annotations
import math
import torch
import torch.nn as nn
import torch.nn.functional as F
from torch.optim.optimizer import Optimizer
try:
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
assert_model_state,
)
except ... | {
"id": "1c49bd6b-88e5-4c63-93c9-07bdb0f64ac3",
"created_at": "2026-08-12 22:40:07.696334+00:00",
"db_md5": "65003311306aad6cda7d6f03df2d3421",
"submitter": "DDanlov",
"github_login": "DDanlov",
"run_id": "cb154f8d-5a5b-4bea-8f2e-40eca07ce1ef",
"tier": "easy",
"dataset_id": "e4",
"status": "failed",
... | null
|
1c4d42d3-66f1-4fd2-9c51-9e47dcca640b | easy | sapient-sapiens | 2026-08-23 07:40:03.883883+00:00 | succeeded | 58ced63ca09398f6843c310ad92a92c1fed233b4669de9eaf4d5d4bcc7701a1e | 19,909 | null | """Approximately 10x-scale adaptive child of the D6 pair/quotient architecture.
The architecture uses the required learned additive token/group/left-relative
representation. Marker-derived field lengths additionally route categorical
digit embeddings into right-aligned lanes; this alternative place alignment is
an ex... | {
"id": "1c4d42d3-66f1-4fd2-9c51-9e47dcca640b",
"created_at": "2026-08-23 07:40:03.883883+00:00",
"db_md5": "e719a7088b2096253e661a56581019bc",
"submitter": "Ertondy",
"github_login": "sapient-sapiens",
"run_id": "5c54dfa6-4d62-4a3e-8d55-0d40943deb9a",
"tier": "easy",
"dataset_id": "e5",
"status": "su... | {
"score": {
"mean_loss": 2.14310588274284,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.009583333333333333
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.146031710064465,
"example_count": 600,
... |
1c5efd67-ec6e-418a-a23e-aede8c9d3c74 | easy | 0Chris5R | 2026-08-01 07:24:29.112448+00:00 | succeeded | afffea9f2e1f130f4bc887b6d9a807270a8d14bbc013c777c9aa16d9f1513eb1 | 23,440 | null | """Shared-digit Neural GPU with late full-objective EMA evaluation.
The model aligns decimal digits by significance, forms a learned low-rank
second-order feature map of the current state, and evolves a two-dimensional
workspace with the original convolutional-GRU equations. Each arithmetic
transition applies the sam... | {
"id": "1c5efd67-ec6e-418a-a23e-aede8c9d3c74",
"created_at": "2026-08-01 07:24:29.112448+00:00",
"db_md5": "a05f77cdb07e4c7e6abef69cd8324275",
"submitter": "Chris ",
"github_login": "0Chris5R",
"run_id": "9d5e151c-d990-40de-a26d-3b4fa49a55d7",
"tier": "easy",
"dataset_id": "e5",
"status": "succeeded"... | {
"score": {
"mean_loss": 2.6344147639227424,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.035000000049670535
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.466576999612988,
"example_count": 600... |
1c63646a-25ee-4ac6-b0df-7354f7341a67 | easy | yashkant | 2026-08-12 04:21:03.308389+00:00 | succeeded | 68938a5bc3a39e8289960f6dd726536195fdf9c8d13a0e22c9e1304c8684d1dd | 31,388 | null | """Round1702 one-projection-per-T Dykstra recursion on the D416 carrier.
Each parsed public-T cycle applies one shared learned smooth projection and
alternates the A/B set code and correction memory across cycles. The immutable
anchor is reinjected on every call; all active cycle outputs are averaged, and
terminal CE... | {
"id": "1c63646a-25ee-4ac6-b0df-7354f7341a67",
"created_at": "2026-08-12 04:21:03.308389+00:00",
"db_md5": "a9eaf9c9fa3858fac12d940ff5a25d6d",
"submitter": "Yash Kant",
"github_login": "yashkant",
"run_id": "0e3b0054-2e9e-48fd-889e-6696c9890086",
"tier": "easy",
"dataset_id": "e2",
"status": "succeed... | {
"score": {
"mean_loss": 3.9135611057281494,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.1575000025331974
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 5.493327617645264,
"example_count": 300,
... |
1c63e5e9-6d48-4791-a61c-5172e5679a99 | easy | yashkant | 2026-08-16 00:58:51.617983+00:00 | succeeded | 6dbfc430384ca1087eaf7fc72e60d7738e186edd8344cbcdfeb995521339e0a9 | 36,391 | null | """Round1790 N-conditional RMS plus mean-free correction crossover.
Both arms retain Round1788 exactly at public N=899 and retain its conditional
RMS-preserving set-code geometry elsewhere. On non-N=899 rows only, the
candidate additionally removes the coordinate mean from each initial and
recurrent Dykstra correctio... | {
"id": "1c63e5e9-6d48-4791-a61c-5172e5679a99",
"created_at": "2026-08-16 00:58:51.617983+00:00",
"db_md5": "3cc74135c7a2c296cbbdaa23f3e7bbeb",
"submitter": "Yash Kant",
"github_login": "yashkant",
"run_id": "cf3aba51-4ad8-48f3-a7d8-8d1534a198ff",
"tier": "easy",
"dataset_id": "e3",
"status": "succeed... | {
"score": {
"mean_loss": 2.1278165082158265,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.011250000158324837
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.1269158489522586,
"example_count": 80... |
1c651a64-3b4a-4e67-9562-485437a78d71 | easy | AFeinfield | 2026-08-28 03:25:33.062417+00:00 | succeeded | f7bac9d1bd9eb9489f2bf0ebc6ba4f47b3d04d59185913117c0f4905d1c96d71 | 10,286 | null | """181: 180 + RESIDUE-SPACE SUPERVISION of F (the literature setting).
Local probe: F_map_acc stayed 0.000 while loss fell — the wide digit
head shortcuts around F. Fix: re-encode provided label digits into a
residue index inside the loss and add CE on F's own-T logits — direct
map supervision, exactly how published g... | {
"id": "1c651a64-3b4a-4e67-9562-485437a78d71",
"created_at": "2026-08-28 03:25:33.062417+00:00",
"db_md5": "982b958b819ec3e4de109a367242d892",
"submitter": "AFeinfield",
"github_login": "AFeinfield",
"run_id": "9dc923dd-0ae8-4ade-92c5-f1b997ce3e92",
"tier": "easy",
"dataset_id": "e6",
"status": "succ... | {
"score": {
"mean_loss": 0.9090563058853149,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.9493243396282196
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 0.8233209848403931,
"example_count": 60,
... |
1c68f4c2-faee-4fb4-a02e-b1e629e33631 | easy | viridale | 2026-08-22 11:04:23.005441+00:00 | failed | 72a960cfe37e74bc786849f6a5c36ac2185438b9a7f8cceeecb67fe0b7745b03 | 116,881 | null | """NV058: compiled exact-candidate executor.
hypothesis: v809 spends almost all failed-state time inside detached candidate
execution, while re-decoding a certified identity recurrent map, rebuilding
fixed categorical values, and constructing zero-valued soft emission graphs.
Compile only those no-gradient... | {
"id": "1c68f4c2-faee-4fb4-a02e-b1e629e33631",
"created_at": "2026-08-22 11:04:23.005441+00:00",
"db_md5": "fb0277be4d2b2397619606f7d6fd0f95",
"submitter": "priormancer",
"github_login": "viridale",
"run_id": "41b2be4d-a461-408b-acd8-146a250970f3",
"tier": "easy",
"dataset_id": "e1",
"status": "faile... | null
|
1c694b77-078c-4f56-b621-670a578eea58 | easy | ddoan | 2026-08-14 21:52:15.130379+00:00 | succeeded | da27b4eb79cdd075f62a754f0267bd10ef945dd9b8a4eecd6b116abb3ddc9517 | 13,997 | null | """Learned transient-cycle clockwork with an attractor-resistant decoder."""
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,... | {
"id": "1c694b77-078c-4f56-b621-670a578eea58",
"created_at": "2026-08-14 21:52:15.130379+00:00",
"db_md5": "c85c04b4f677d3053cb262c1be6cf632",
"submitter": "Doug Doan",
"github_login": "ddoan",
"run_id": "76106fe3-8869-4401-b277-8ef80333ca05",
"tier": "easy",
"dataset_id": "e1",
"status": "succeeded"... | {
"score": {
"mean_loss": 1.8644787073135376,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.07999999821186066
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 1.8030823469161987,
"example_count": 100... |
1c69fcc7-ddb8-4891-810f-f4991fcda651 | easy | TZZheng | 2026-08-07 05:17:33.881068+00:00 | succeeded | 5522232b425e827573ac2e7e61e9c244cfede8ec50972edbd6ccdced42389c63 | 12,525 | null | """Learned recurrent modular-transition model.
The mutable digit state is initialized from x. A learned square-oriented
self-mixing network and an N-conditioned reduction network form one transition,
and exactly the same transition parameters are reused T times. No arithmetic
operation is hard-coded; only the public p... | {
"id": "1c69fcc7-ddb8-4891-810f-f4991fcda651",
"created_at": "2026-08-07 05:17:33.881068+00:00",
"db_md5": "6a169a69739e647fb20a5aefaf07fe28",
"submitter": "TZZheng",
"github_login": "TZZheng",
"run_id": "3bc5fde0-57b5-439d-a15d-e6cba8c36d8c",
"tier": "easy",
"dataset_id": "e1",
"status": "succeeded"... | {
"score": {
"mean_loss": 9.857700109481812,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.04333333298563957
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 12.460246086120605,
"example_count": 100,... |
1c717c39-463f-4f2f-9608-c356cddfceac | easy | himalalps | 2026-08-09 12:23:23.998865+00:00 | succeeded | da9f5ec605d6070176963339426f39c1d0d07fe6cc690c48fb021838acde31fc | 81,115 | null | """Algorithmically constrained latent recurrence for One Layer Deeper.
One masked cross-attention pass creates two independent representations:
``state_0`` from X only and an immutable context from N only. T is excluded
from both representations and is used solely to choose how many times the
learned recurrent MLP tr... | {
"id": "1c717c39-463f-4f2f-9608-c356cddfceac",
"created_at": "2026-08-09 12:23:23.998865+00:00",
"db_md5": "0c65690084b54fa2371809e4f31caccf",
"submitter": "Haoyu Tang",
"github_login": "himalalps",
"run_id": "3c5781f7-3ffc-466f-bfe9-1c63841d658b",
"tier": "easy",
"dataset_id": "e5",
"status": "succe... | {
"score": {
"mean_loss": 2.19237177258542,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.009583333233992259
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.1794540882110596,
"example_count": 600,... |
1c78521f-4828-4ba9-a0c7-ec0d6822b293 | easy | viridale | 2026-08-02 07:41:58.121519+00:00 | succeeded | 5b4131612fe0e2969b3220fe754d4a8445cfd33d641b0da96f20656b20442ed9 | 27,336 | null | """Sequence-global complete-program evidence through TokenLossBatch.
hypothesis: faithful runner-order instrumentation rejects v084/v085's clip
diagnosis: factors remain exact, but fixed-N training cannot distinguish
r^2 from 900*r^2 modulo N=899. Adding the same saturated evidence update to
exact aliases ... | {
"id": "1c78521f-4828-4ba9-a0c7-ec0d6822b293",
"created_at": "2026-08-02 07:41:58.121519+00:00",
"db_md5": "ef7b3e15afc20d6dae7581eb8ca2af7f",
"submitter": "priormancer",
"github_login": "viridale",
"run_id": "7294c3b2-ef79-4417-9dfa-11657e93b9e0",
"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,
"... |
1c79303b-f999-4386-93cb-31a05865f4d7 | easy | viridale | 2026-08-06 10:18:41.488713+00:00 | succeeded | a9489938720e8949627831883d4a0b9cd0545aea4bb9530a592e0e1f16edad01 | 9,463 | null | """Append-only vertical recurrent trace with attendable prior states.
hypothesis: overwriting one recurrent state destroys useful partial computations;
retaining every within-example repeat state as learned attention keys and values
should give later repeats a direct retrieval and gradient path.
axis: arch
tar... | {
"id": "1c79303b-f999-4386-93cb-31a05865f4d7",
"created_at": "2026-08-06 10:18:41.488713+00:00",
"db_md5": "1ea53afc30b75ca185806fe8d3e404f5",
"submitter": "priormancer",
"github_login": "viridale",
"run_id": "3759b6a5-97dc-4cb3-8b27-03ba371d4152",
"tier": "easy",
"dataset_id": "e1",
"status": "succe... | {
"score": {
"mean_loss": 2.1795449256896973,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0833333320915699
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.1914596557617188,
"example_count": 100,... |
1c817490-a347-4f31-80ee-5ab0c5b347db | easy | DESU-CLUB | 2026-08-23 23:35:53.819109+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": "1c817490-a347-4f31-80ee-5ab0c5b347db",
"created_at": "2026-08-23 23:35:53.819109+00:00",
"db_md5": "eaac69fe345f82973fcf392c504a40ff",
"submitter": "Low Keng Hoong, Warren",
"github_login": "DESU-CLUB",
"run_id": "bfd9c39b-cc03-4652-99d6-cbfc914675ea",
"tier": "easy",
"dataset_id": "e6",
"sta... | {
"score": {
"mean_loss": 4.529300093650818,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.011711712228134274
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 6.157750129699707,
"example_count": 60,
... |
1c82952c-2a37-473e-b8be-14b16eae3265 | easy | pepinu | 2026-08-26 17:57:14.567952+00:00 | succeeded | 74307ba76d035d71e8330e51db839434ca24ae8d3efbab3f4b12dd2ec11dd273 | 46,270 | 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": "1c82952c-2a37-473e-b8be-14b16eae3265",
"created_at": "2026-08-26 17:57:14.567952+00:00",
"db_md5": "d7f34595d7f996ce3ab1db374501668a",
"submitter": "pepinu",
"github_login": "pepinu",
"run_id": "c6097f7c-6855-4360-90d0-2970a61196ed",
"tier": "easy",
"dataset_id": "e10",
"status": "succeeded",... | {
"score": {
"mean_loss": 0.013974567409604788,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 1.0
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 0.013897448778152466,
"example_count": 125,
... |
1c89496e-e02e-4908-9d8c-48c59614a52f | easy | jordanrubin | 2026-08-08 10:00:06.231515+00:00 | succeeded | 078cd8887c504bdcfcf580d442de465beb5b3a0ea5e1726a6c95fd2267731c91 | 16,615 | null | """Bidirectional transformer, D=256 x 3 blocks, grokking-tuned AdamW.
Design notes:
- Inputs are padded to config.max_seq_len inside forward and logits sliced
back, so every training batch presents one static shape (compile/cudagraph
friendly, and uniform kernel shapes even in eager).
- Attention mask gets an iden... | {
"id": "1c89496e-e02e-4908-9d8c-48c59614a52f",
"created_at": "2026-08-08 10:00:06.231515+00:00",
"db_md5": "1f66d0c9d8b2addfecdef9a9aa694dae",
"submitter": "Jordan Rubin",
"github_login": "jordanrubin",
"run_id": "6903390a-f762-4afa-bf02-5a81b7fee88f",
"tier": "easy",
"dataset_id": "e5",
"status": "s... | {
"score": {
"mean_loss": 2.179996594796276,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.00458333323088785
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.182715232864089,
"example_count": 600,
... |
1c89e936-da17-432f-b937-6166be84d1e9 | easy | newjordan | 2026-08-09 16:46:47.099834+00:00 | succeeded | 72f7c275f49c94dd8984da15a4db9c67da644307458dfe8ecf639899b63d644f | 93,178 | null | """Helix-Crawler — a recursive-weight-shared "crawler" ALU for modular squaring.
The task is r_{i+1} = r_i^2 mod n, chained T times: a *recurrence* over a single
operator. This build embodies the crawler / Frugendorff lineage (spark:~/sota_crawler)
mapped onto that structure. n never enters as a table row (unseen modu... | {
"id": "1c89e936-da17-432f-b937-6166be84d1e9",
"created_at": "2026-08-09 16:46:47.099834+00:00",
"db_md5": "eb1cfe41b2cc660a4a5f78440dba6c1d",
"submitter": "Frosty40",
"github_login": "newjordan",
"run_id": "d3e29ac8-df93-4cd4-91f7-9e5058138b9a",
"tier": "easy",
"dataset_id": "e3",
"status": "succeed... | {
"score": {
"mean_loss": 2.828032845320881e-06,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 1.0
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.828571163734796e-06,
"example_count": 800,
... |
1c90f91d-6d1d-4368-b301-2c30963b8c06 | easy | jordanrubin | 2026-08-13 18:18:06.580979+00:00 | succeeded | e2245c74dfcfe7d00f8aeee61890a07c07a0d89977c245e3844890ce7f77b424 | 27,175 | 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": "1c90f91d-6d1d-4368-b301-2c30963b8c06",
"created_at": "2026-08-13 18:18:06.580979+00:00",
"db_md5": "3f710550312ea469d13b008d83980242",
"submitter": "Jordan Rubin",
"github_login": "jordanrubin",
"run_id": "295a970e-7970-4198-a0c4-67c03749f54d",
"tier": "easy",
"dataset_id": "e5",
"status": "s... | {
"score": {
"mean_loss": 2.851825484432199,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.027916665577019256
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.241096103137919,
"example_count": 600,... |
1c937e13-7ca0-4761-8dab-392660d7bc18 | easy | kushasareen | 2026-08-12 02:12:38.251432+00:00 | succeeded | 247ada0ffe8d033924f089e14816a48f6888db846afbbeb666f64e0a4003348e | 6,916 | null | """v2 capacity descent: d_model 184, lr 2.8e-3 (1/width scaling).
T-conditioned recurrent Transformer with the measured stability fixes:
- 2 untied pre-blocks (pre-norm), 1 untied post-block, tied loop block
- prompt re-injection every loop iteration (gradient + information path)
- UT-style double post-norm inside the... | {
"id": "1c937e13-7ca0-4761-8dab-392660d7bc18",
"created_at": "2026-08-12 02:12:38.251432+00:00",
"db_md5": "5a74d1724963dde28798e4e6f70a160c",
"submitter": "Kusha Sareen",
"github_login": "kushasareen",
"run_id": "cfe0367c-35f8-47b1-b76f-9a2d98bdc58d",
"tier": "easy",
"dataset_id": "e5",
"status": "s... | {
"score": {
"mean_loss": 8.94637414716843,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0054166666790843015
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 8.87309295927997,
"example_count": 600,
... |
1c946a83-22b1-4b0e-ba44-81a632531b36 | easy | arthurfeeney | 2026-08-07 07:34:17.681001+00:00 | succeeded | bf6186a85a1363eafb06221e59af0f1a1ba4eafc5ff838276f68c69ef173dbe3 | 12,918 | null | r"""
Notes
1. goal is basically learn y = G(x_0, T, N).
G is always a recurrence `x_t+1 = f(x_t, t) mod N`.
Shouldn't be possible in general since f isn't determinable from a dataset...
I.e., multiple f can generate the same training dataset.
2. can't really use any info on structure of f...
"""
from __future... | {
"id": "1c946a83-22b1-4b0e-ba44-81a632531b36",
"created_at": "2026-08-07 07:34:17.681001+00:00",
"db_md5": "0ab7aacbaf2a583907906daba0cff7aa",
"submitter": "Arthur",
"github_login": "arthurfeeney",
"run_id": "c7e7c252-79e5-402d-8947-68e91dadfbe5",
"tier": "easy",
"dataset_id": "e1",
"status": "succee... | {
"score": {
"mean_loss": 8.998088359832764,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.04666666593402624
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 9.78657341003418,
"example_count": 100,
... |
1c96de38-e000-4231-a413-29d19a0ef7c1 | easy | adwhit | 2026-08-14 08:15:18.781073+00:00 | succeeded | 265436df2f8f5ebab03d48ddb4d2ecc2a4cb9bb0b1bd67be9148be8224ce7ee8 | 17,774 | null | """GaussNet: additive-character recurrent model for One Layer Deeper.
Structure (all parameters learned from random init):
- Learned parse: cross-attention queries extract LSB-first digit registers
for x, copy-supervised from the prompt's own tokens via the auxiliary
channel (self-supervision, no task math).... | {
"id": "1c96de38-e000-4231-a413-29d19a0ef7c1",
"created_at": "2026-08-14 08:15:18.781073+00:00",
"db_md5": "37f6d502b050a37f4faae457446910e3",
"submitter": "adwhit",
"github_login": "adwhit",
"run_id": "58394104-ad2f-4e36-a964-759ae1297f10",
"tier": "easy",
"dataset_id": "e2",
"status": "succeeded",
... | {
"score": {
"mean_loss": 2.4470468759536743,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.449591875076294,
"example_count": 300,
"exac... |
1ca094db-4185-4efc-897a-f38112a874b5 | easy | khushidahi | 2026-08-21 21:11:33.803902+00:00 | succeeded | 1487dd42e751358af1ff205598857d8772f74326add1bf070d0038e389ca2c1c | 39,548 | null | """R365: balanced-gradient optimization of R318's polynomial root family.
R318 is retained as the architectural control: a low-rank learned quadratic
product tape, a learned N-conditioned causal reducer, categorical recurrent
state, and evaluator-final-label-only supervision. The sole experimental
change is Balanced ... | {
"id": "1ca094db-4185-4efc-897a-f38112a874b5",
"created_at": "2026-08-21 21:11:33.803902+00:00",
"db_md5": "ff7a4bad628cadf17e10880d9bb65720",
"submitter": "khushidahi",
"github_login": "khushidahi",
"run_id": "32b05b64-7d96-4c23-a04b-3778f2431147",
"tier": "easy",
"dataset_id": "e3",
"status": "succ... | {
"score": {
"mean_loss": 2.231699767633845,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.006875000000000001
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.225081527238139,
"example_count": 800,... |
1ca63371-fe26-4d51-95fa-c27234af4db3 | easy | shirvani-jr | 2026-08-27 00:04:54.983363+00:00 | succeeded | 6ad27736f670a641a4f54f9c72c67ce54a848045d23897c38127ec129f43c687 | 26,108 | null | """Adaptive learned discrete recurrent fabric for One Layer Deeper.
A gradient-trained, operation-free recurrent computational machine. The
model learns, from the competition's endpoint supervision: how to represent
the prompt, what to store in a bank of categorical registers, a reusable
transition circuit applied se... | {
"id": "1ca63371-fe26-4d51-95fa-c27234af4db3",
"created_at": "2026-08-27 00:04:54.983363+00:00",
"db_md5": "e075ffc25f31ede242c3f762c9723ee1",
"submitter": "Ali",
"github_login": "shirvani-jr",
"run_id": "edc716e3-d2f9-4036-abe5-919880f07ac9",
"tier": "easy",
"dataset_id": "e1",
"status": "succeeded"... | {
"score": {
"mean_loss": 2.162655830383301,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.023333333432674408
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.1527633666992188,
"example_count": 100... |
1ca7d4b2-2af4-403a-a268-f36b451afb7e | easy | 0Chris5R | 2026-08-01 07:15:52.797738+00:00 | succeeded | 07bb9cec4fc3917a6a62fb9d750d1801bb1529956dece7c7e208c44f7acd2868 | 11,547 | null | """Recurrent workspace with learned second-order relational memory.
The tensor-memory head exposes generic pairwise token interactions to a learned
recurrent model. It does not select fields or encode any arithmetic operation.
"""
from __future__ import annotations
import math
import torch
import torch.nn.functiona... | {
"id": "1ca7d4b2-2af4-403a-a268-f36b451afb7e",
"created_at": "2026-08-01 07:15:52.797738+00:00",
"db_md5": "bc21e43cd95098fed494b3d5915e80bf",
"submitter": "Chris ",
"github_login": "0Chris5R",
"run_id": "9db16dd4-7017-4024-b2f2-8246c4417b1b",
"tier": "easy",
"dataset_id": "e5",
"status": "succeeded"... | {
"score": {
"mean_loss": 7.709949099039127,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.009583333358168602
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 7.696248184939671,
"example_count": 600,... |
1cafd436-6c69-43ba-87b5-267ebd0b6be9 | easy | AFeinfield | 2026-08-26 22:11:01.795482+00:00 | succeeded | bfc563675c0ea37dd500960cd024e9d0e9dc45ce4c45807ecdabe0538d337c26 | 11,683 | null | """138: 136 + lr 6e-3 (hot-LR retest on the non-transformer chassis;
130's negative was a different architecture
passes per squaring, bidirectional carry/borrow sweeps).
135 validated the coordinate system decisively: escape in 13s on m6 and
2x-earlier escape on m3 at 129K params — then stalled at a new lower
plateau ... | {
"id": "1cafd436-6c69-43ba-87b5-267ebd0b6be9",
"created_at": "2026-08-26 22:11:01.795482+00:00",
"db_md5": "f935b83f854776364567fa6d81527d48",
"submitter": "AFeinfield",
"github_login": "AFeinfield",
"run_id": "20b08caf-f14b-4dda-9c9b-26edbed566f8",
"tier": "easy",
"dataset_id": "e5",
"status": "succ... | {
"score": {
"mean_loss": 2.619972972479173,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.005
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.711183371565267,
"example_count": 600,
"exa... |
1cb48665-c9fe-417b-a7ba-20778dd414ef | easy | khushidahi | 2026-08-12 21:10:17.763163+00:00 | succeeded | f48f46a2c5e48d7d6ea1c5c3155b8eafcc26670c1bc80aeea807db27b251fcca | 16,710 | 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": "1cb48665-c9fe-417b-a7ba-20778dd414ef",
"created_at": "2026-08-12 21:10:17.763163+00:00",
"db_md5": "bd5e36231dcb0d22a724bd864a963b90",
"submitter": "khushidahi",
"github_login": "khushidahi",
"run_id": "f14c4f9d-e0f6-4a59-a800-bedb49b33a2b",
"tier": "easy",
"dataset_id": "e5",
"status": "succ... | {
"score": {
"mean_loss": 4.248066663742065,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.007083333563059568
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 4.2259063720703125,
"example_count": 600... |
1cb90683-bb31-47e4-95d1-90cb5bc3f759 | easy | benjaminW2025 | 2026-08-18 00:19:47.331002+00:00 | succeeded | 3e9791093272569c751b274682bdd6f5fb07f82a313f7f09a4575f734ad3e481 | 11,144 | null | """Basic single-pass Transformer with PyTorch AdamW."""
from __future__ import annotations
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
assert_model_state,
)
D_MODEL = 256
NUM_HEADS = 4
... | {
"id": "1cb90683-bb31-47e4-95d1-90cb5bc3f759",
"created_at": "2026-08-18 00:19:47.331002+00:00",
"db_md5": "29614049a8555a91c034670c8ec9b7da",
"submitter": "benjawesome",
"github_login": "benjaminW2025",
"run_id": "5534b1d5-d56e-422e-9101-acf7700d01d2",
"tier": "easy",
"dataset_id": "e5",
"status": "... | {
"score": {
"mean_loss": 2.2471350595378814,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.012500000037252903
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.2396007557086346,
"example_count": 60... |
1cbb31ef-cd9a-44a6-989c-2e40a4bcfd31 | easy | himalalps | 2026-08-07 02:27:11.093008+00:00 | succeeded | 5a1ca372493a34bf311ad1b651d6e26cc9b7a9df8a4e8b1645bddfb0e14a0a75 | 65,432 | null | """Algorithmically constrained latent recurrence for One Layer Deeper.
One masked cross-attention pass creates two independent representations:
``state_0`` from X only and an immutable context from N only. T is excluded
from both representations and is used solely to choose how many times the
learned recurrent MLP tr... | {
"id": "1cbb31ef-cd9a-44a6-989c-2e40a4bcfd31",
"created_at": "2026-08-07 02:27:11.093008+00:00",
"db_md5": "91332ebae51b8a4518cfc824bd6bbf97",
"submitter": "Haoyu Tang",
"github_login": "himalalps",
"run_id": "0d33383b-44bd-4d7b-9155-acdafb08f5cd",
"tier": "easy",
"dataset_id": "e5",
"status": "succe... | {
"score": {
"mean_loss": 2.1936165700154797,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.01249999988824129
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.196824789047241,
"example_count": 600,... |
1cbc08db-5617-4d6e-8cc2-278ebf65e45f | easy | erdavis0 | 2026-08-22 07:01:09.563308+00:00 | succeeded | 598ed87f6b0dc7d2ed53a90de6130bdd91679d4793de9d6fb753bff22d2ae00d | 5,871 | null | """Bidirectional gated carry scan with learned decimal-state channels."""
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
... | {
"id": "1cbc08db-5617-4d6e-8cc2-278ebf65e45f",
"created_at": "2026-08-22 07:01:09.563308+00:00",
"db_md5": "39b9a5496bba32333c8bb77c0f49e9ba",
"submitter": "Ethan",
"github_login": "erdavis0",
"run_id": "24e264b9-7216-40e1-a070-7b14691217a4",
"tier": "easy",
"dataset_id": "e4",
"status": "succeeded",... | {
"score": {
"mean_loss": 4.942609548568726,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.002187499951105565
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 5.116909980773926,
"example_count": 1200... |
1cc1de8d-c96d-4544-bb07-b742951b4085 | easy | rteehas | 2026-08-19 08:26:55.512678+00:00 | succeeded | 89938d6e70952c6215586e8bebe56eb591140b567b9149d5ca797584302788db | 41,409 | null | """Anonymous marginal-program submission.
Mechanism
---------
The model is a differentiable probabilistic program. Its forward pass
computes, exactly, the answer-token predictions of a small structured latent
hypothesis space:
(prompt segmentation = which token types head the three fields)
x (field-role permutat... | {
"id": "1cc1de8d-c96d-4544-bb07-b742951b4085",
"created_at": "2026-08-19 08:26:55.512678+00:00",
"db_md5": "f8da87d40944b0aee64044f85b4c7e51",
"submitter": "Ryan Teehan",
"github_login": "rteehas",
"run_id": "2054184d-b058-4436-9368-ba2a8d745bb7",
"tier": "easy",
"dataset_id": "e5",
"status": "succee... | {
"score": {
"mean_loss": 7.354689874269984,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0020833333457509675
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 7.399971763114758,
"example_count": 600... |
1ccf0051-0df9-4d49-ad2f-5e4d1639cd29 | easy | DDanlov | 2026-08-15 14:54:30.997773+00:00 | succeeded | 252d1832bfe0ff05c62a9c61eb1c83bac7a30c7a6e9a609d6948a00bfd97ae05 | 19,147 | null | """
Official Submission for One Layer Deeper Challenge
Architecture: Compound 4-Layer Pure No-Norm No-Skip DKS Reasoning Pipeline:
- 1 Prefix Layer (prompt binding)
- 2 Weight-Tied Recurrent Core Layers (modular squaring dynamics unrolled T times)
- 1 Suffix Layer (place-value digit decoding)
- dim=512, 16 heads, d_ff=... | {
"id": "1ccf0051-0df9-4d49-ad2f-5e4d1639cd29",
"created_at": "2026-08-15 14:54:30.997773+00:00",
"db_md5": "14957b49613d11cb149365c9a66187c0",
"submitter": "DDanlov",
"github_login": "DDanlov",
"run_id": "8ff1ecf8-bab7-43e4-831a-fa2d158309d5",
"tier": "easy",
"dataset_id": "e1",
"status": "succeeded"... | {
"score": {
"mean_loss": 8.074770512408538,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.050000000173846884
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 5.595275893140195,
"example_count": 100,... |
1cd15a7e-1399-48ab-af60-0ab255e50c98 | easy | himalalps | 2026-08-06 07:30:44.500597+00:00 | succeeded | bfb095af1f6acd16462d80395ef4bc55b878402f7a6b2ba4d2d424da9684af2c | 22,874 | null | """Algorithmically constrained latent recurrence for One Layer Deeper.
One masked cross-attention pass creates two independent representations:
``state_0`` from X only and an immutable context from N only. T is excluded
from both representations and is used solely to choose how many times the
learned recurrent MLP tr... | {
"id": "1cd15a7e-1399-48ab-af60-0ab255e50c98",
"created_at": "2026-08-06 07:30:44.500597+00:00",
"db_md5": "4131cdd151c02aff53585593b417527a",
"submitter": "Haoyu Tang",
"github_login": "himalalps",
"run_id": "3ffe8969-b103-462e-99a2-71bfc7441582",
"tier": "easy",
"dataset_id": "e2",
"status": "succe... | {
"score": {
"mean_loss": 8.398543357849121,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.011666667181998491
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 8.6165189743042,
"example_count": 300,
... |
1cd8acf8-0fea-4b34-a5b0-22f5c4de4b99 | easy | kushasareen | 2026-08-10 21:07:43.975601+00:00 | succeeded | b6bb3399161266d22f6ddfc9acde1f19955e84238d6b9c5f1316515de57a9e71 | 5,775 | null | """T-conditioned recurrent Transformer.
A tied-weight block is applied once per requested squaring step, so compute
scales with T and the model can extrapolate to depths never seen in training.
Untied pre-blocks encode the prompt and an untied post-block decodes the
answer digits.
"""
from __future__ import annotatio... | {
"id": "1cd8acf8-0fea-4b34-a5b0-22f5c4de4b99",
"created_at": "2026-08-10 21:07:43.975601+00:00",
"db_md5": "bf42d5a9f20cb7599536b63045f7342b",
"submitter": "Kusha Sareen",
"github_login": "kushasareen",
"run_id": "86774c00-4347-42be-9a03-1652094cf19c",
"tier": "easy",
"dataset_id": "e1",
"status": "s... | {
"score": {
"mean_loss": 2.5831241607666016,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.018333333544433117
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.18851375579834,
"example_count": 100,... |
1cda2490-c335-4fd7-bdb2-615fa8060c2d | easy | DDanlov | 2026-08-17 07:54:39.913777+00:00 | succeeded | 13f3f950251e49ff4332e11ebf512d179a4c66422710953d3be7725445e4401e | 28,291 | null | """
Scalable Recurrent Deliberation Transformer (ScalableDEPHNet)
============================================================
Architecture:
- Pure No-Norm: Single unit RMS normalization strictly isolated to the input token embedding layer.
Zero LayerNorm / RMSNorm inside recurrent deliberation blocks.
- Pure No-Skip... | {
"id": "1cda2490-c335-4fd7-bdb2-615fa8060c2d",
"created_at": "2026-08-17 07:54:39.913777+00:00",
"db_md5": "b93a777599800f02ee2e5325d67a5d39",
"submitter": "DDanlov",
"github_login": "DDanlov",
"run_id": "b8483577-106d-4b86-9353-9cd9e69bb1e5",
"tier": "easy",
"dataset_id": "e1",
"status": "succeeded"... | {
"score": {
"mean_loss": 1.900140730222859,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.01
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 1.9054464101791382,
"example_count": 100,
"exa... |
1cdda8a2-c0df-4f87-b320-536c768a0657 | easy | karanganesan | 2026-08-04 00:42:19.949478+00:00 | succeeded | c0fea402866a5402f22235da7527f9a2d2aff2db4d0c6e4be230a6b950cc24da | 11,766 | null | """Parametric looped-transformer family (P1).
One weight-tied transformer block applied k times in latent space. Config
flags cover four P1 families with one file:
looped recall=0 gated=0 tfilm=0 plain weight-tied loop
looped-recall recall=1 re-inject the input embedding each
... | {
"id": "1cdda8a2-c0df-4f87-b320-536c768a0657",
"created_at": "2026-08-04 00:42:19.949478+00:00",
"db_md5": "46e0560c11000885c350106bc692db4c",
"submitter": "Karan Ganesan",
"github_login": "karanganesan",
"run_id": "96d74265-6978-42f6-9df1-a48c999e1d0a",
"tier": "easy",
"dataset_id": "e1",
"status": ... | {
"score": {
"mean_loss": 2.3193109035491943,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0833333320915699
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.437121868133545,
"example_count": 100,
... |
1ce10bb0-4eaa-4706-bb43-da4a1495d194 | easy | ehonig | 2026-08-13 22:54:43.210472+00:00 | succeeded | 7eeea52578ae2bd443e3739bd1d98133ffa1fdcc87cc7d175753d015af4c3b26 | 45,086 | null | """DigitLoop — a place-value recurrent model for repeated modular squaring.
The task is ``y = x^(2^T) mod N`` with the answer read back tail aligned, one
decimal place value per token position from the right. A stack of attention and
MLP layers has no primitive for the two operations this needs — a partial-product
co... | {
"id": "1ce10bb0-4eaa-4706-bb43-da4a1495d194",
"created_at": "2026-08-13 22:54:43.210472+00:00",
"db_md5": "8b1b4231fd42113155c7f0d6f6018a1c",
"submitter": "Edouardo Honig",
"github_login": "ehonig",
"run_id": "d93c09e4-2420-4c7e-babe-7ad742b41a3a",
"tier": "easy",
"dataset_id": "e1",
"status": "succ... | {
"score": {
"mean_loss": 5.202166318893433,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0566666666418314
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 5.340229034423828,
"example_count": 100,
... |
1ce53d3d-0dee-4334-b8af-cb4b93041a1c | easy | jordanrubin | 2026-08-08 09:44:58.288955+00:00 | succeeded | 078cd8887c504bdcfcf580d442de465beb5b3a0ea5e1726a6c95fd2267731c91 | 16,615 | null | """Bidirectional transformer, D=256 x 3 blocks, grokking-tuned AdamW.
Design notes:
- Inputs are padded to config.max_seq_len inside forward and logits sliced
back, so every training batch presents one static shape (compile/cudagraph
friendly, and uniform kernel shapes even in eager).
- Attention mask gets an iden... | {
"id": "1ce53d3d-0dee-4334-b8af-cb4b93041a1c",
"created_at": "2026-08-08 09:44:58.288955+00:00",
"db_md5": "1f66d0c9d8b2addfecdef9a9aa694dae",
"submitter": "Jordan Rubin",
"github_login": "jordanrubin",
"run_id": "db896994-143d-4b0f-8167-03fb1f63fdc9",
"tier": "easy",
"dataset_id": "e2",
"status": "s... | {
"score": {
"mean_loss": 3.7342535797066043,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.21249999447415271
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 6.059015967981602,
"example_count": 300,... |
1ce91dbb-97f5-4e9c-930b-96fbc5bc67ef | easy | sapient-sapiens | 2026-08-21 15:11:45.301940+00:00 | succeeded | dd38f61917056c6a0d251cfa3774db36ea44984868eb074b64e7dea53f2dde3c | 27,459 | null | """Parameterized fixed-K recurrent candidate for the additive sweep.
Knobs at the top are the only intended experimental variables. T remains an
ordinary prompt field and is never parsed into a loop count.
"""
from __future__ import annotations
import math
import time
import torch
import torch.nn.functional as F
fr... | {
"id": "1ce91dbb-97f5-4e9c-930b-96fbc5bc67ef",
"created_at": "2026-08-21 15:11:45.301940+00:00",
"db_md5": "02742984b457ef51e4d4fbc04ad75fc9",
"submitter": "Ertondy",
"github_login": "sapient-sapiens",
"run_id": "a7a64287-8744-433a-81b2-8a800bb23cbf",
"tier": "easy",
"dataset_id": "e5",
"status": "su... | {
"score": {
"mean_loss": 2.3609468442836476,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0075
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.361593666498971,
"example_count": 600,
"e... |
1cef36a1-2d04-4a68-a3cb-625e94815a6d | easy | rougee | 2026-08-25 04:09:24.797713+00:00 | succeeded | be5fd80156dbb71b6ecaf5c48bd82b47763fc3de0c85dcb2f883b1023265aae4 | 13,829 | null | """Official Submittable Champion for One Layer Deeper (Contextual Prelude + Stacked Quadratic Core).
State-of-the-Art Recurrent Architecture:
- Prelude: Causal RoPE Multi-Head Self-Attention Contextual Encoder
- Metric Geometry: Dynamic Skew-Symmetric Lie Group Matrix Exponential W_orth = exp(S - S^T) from Causal Prom... | {
"id": "1cef36a1-2d04-4a68-a3cb-625e94815a6d",
"created_at": "2026-08-25 04:09:24.797713+00:00",
"db_md5": "36af48eb167a1685353bed98710b84a7",
"submitter": "Rougee",
"github_login": "rougee",
"run_id": "c756a4d2-b300-4a58-89d4-489bd14cfa5a",
"tier": "easy",
"dataset_id": "e1",
"status": "succeeded",
... | {
"score": {
"mean_loss": 19.545124053955078,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.03333333320915699
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 17.787857055664062,
"example_count": 100... |
1cfa42ed-a7fb-4850-a8f1-3750c750846e | easy | EyimofeA | 2026-08-16 23:50:41.751469+00:00 | succeeded | 8a18df9250481592776408f520c815627dc8758b3848b9fb668a5ddebebf5253 | 10,907 | null | """Binary fused work-state recurrence: width 256 + tuned flattened Muon.
The model receives exact binary representations of evaluator-provided x and N,
then learns the complete transition from final labels only. It has no explicit
squarer, reducer, arithmetic trace, or task oracle.
"""
from __future__ import annotati... | {
"id": "1cfa42ed-a7fb-4850-a8f1-3750c750846e",
"created_at": "2026-08-16 23:50:41.751469+00:00",
"db_md5": "fcfa2179c5c24d8f99e9c8337749a82b",
"submitter": "mof",
"github_login": "EyimofeA",
"run_id": "0b8b09b9-97bd-43aa-9a88-6d0ea553ed6f",
"tier": "easy",
"dataset_id": "e5",
"status": "succeeded",
... | {
"score": {
"mean_loss": 35.921655704078105,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.005416666691501935
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 35.78475294840175,
"example_count": 600... |
1cff6ec1-0cf3-4167-8c00-95225a172116 | easy | lpbb | 2026-08-19 11:35:44.988775+00:00 | succeeded | 84f7ca58abdd79dd823f9b44474502b5c3dedb4e64cdb166786feafdcf359410 | 9,095 | null | """S0 + RoPE + abacus embeddings + dropout, with a looped block: instead of
`num_layers` independent layers, a block of `num_layers` unique layers is run
`num_repeats` times with shared weights (e.g. 12 layers x 4 repeats = depth 48
but only 12 layers' worth of parameters). Weight tying across repeats forces
the block ... | {
"id": "1cff6ec1-0cf3-4167-8c00-95225a172116",
"created_at": "2026-08-19 11:35:44.988775+00:00",
"db_md5": "ab81a6b15e3f0dc0b8310baaccbff910",
"submitter": "Lpbb",
"github_login": "lpbb",
"run_id": "6df28d1f-570d-4cb5-a3ec-36ad296ed3f8",
"tier": "easy",
"dataset_id": "e1",
"status": "succeeded",
"s... | {
"score": {
"mean_loss": 2.56307852268219,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.04833333380520344
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 3.1085009574890137,
"example_count": 100,
... |
1d06b767-e6b9-4d0d-b37a-d538dae7a9e1 | easy | liam-gb | 2026-08-14 00:05:59.638873+00:00 | succeeded | 1d746e3ae8d56e5df34ccf93f23d12fd3fcbc2567f189b9b9955d782eab23df6 | 15,758 | null | """rns_pm_gather_ct_widet_b46rnore: 45-prime bank + per-prime temperature
+ tables at 3e-2 + reuse OFF -- every lever that has won tonight, combined.
Base: 45-prime bank + per-prime temperature + tables at lr 3e-2.
Base: widet + per-prime temperature + 45-prime bank.
The wide readout is what makes long answers reach... | {
"id": "1d06b767-e6b9-4d0d-b37a-d538dae7a9e1",
"created_at": "2026-08-14 00:05:59.638873+00:00",
"db_md5": "b8e9c96ce32883e9498e263123a1efcd",
"submitter": "liam-gb",
"github_login": "liam-gb",
"run_id": "af201d76-a10b-4792-ac1b-bed0494dcb05",
"tier": "easy",
"dataset_id": "e5",
"status": "succeeded"... | {
"score": {
"mean_loss": 7.266389372850861,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.06833333348855376
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 7.810051441192627,
"example_count": 600,
... |
1d08c08e-7c03-43ed-8f75-13891443dbe9 | easy | sapient-sapiens | 2026-08-04 16:06:37.441787+00:00 | succeeded | 2e9ef7c86a5507a567d20b7fdca2ce258901df0e47a01bd1fd2575e05ef86983 | 18,008 | null | """Weight-tied recurrent-depth Transformer with progressive + hold losses.
Phase 2: prelude / recalled tied core / coda, differentiable hold gate,
fp32 core under disabled autocast, random train depth, K=72 at eval.
Extra depth readouts travel through the auxiliary return value into
token_training_loss. Wall-clock Ada... | {
"id": "1d08c08e-7c03-43ed-8f75-13891443dbe9",
"created_at": "2026-08-04 16:06:37.441787+00:00",
"db_md5": "aaf3156753b39155c0ffa8e9e35fd63a",
"submitter": "Ertondy",
"github_login": "sapient-sapiens",
"run_id": "82b65eb2-e630-426e-960e-0f3d97ce7362",
"tier": "easy",
"dataset_id": "e5",
"status": "su... | {
"score": {
"mean_loss": 2.1888821207572944,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0020833333457509675
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.1953096695812295,
"example_count": 6... |
1d0a2905-ed7a-48b1-87a4-5a710bf7a035 | easy | lpbb | 2026-08-12 08:28:24.880812+00:00 | succeeded | f60f6021bb5e247f983a9a1b971ff72df71e09e97602529ffee5b95b1e6f125d | 24,142 | 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": "1d0a2905-ed7a-48b1-87a4-5a710bf7a035",
"created_at": "2026-08-12 08:28:24.880812+00:00",
"db_md5": "ffc44ff5cc3743e77299143eaf88f930",
"submitter": "Lpbb",
"github_login": "lpbb",
"run_id": "8a1c8dce-7940-4109-b7cb-01acba67b289",
"tier": "easy",
"dataset_id": "e1",
"status": "succeeded",
"s... | {
"score": {
"mean_loss": 0.8363014310598373,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.8616666495800018
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 0.11742100119590759,
"example_count": 100... |
1d1062b6-5700-4274-9330-3eecdd7031e8 | easy | TheYkk | 2026-08-04 13:56:58.757310+00:00 | succeeded | d8433be4033e1fbd5939dfa50e2c261d42a43dc2ff22b87b166b579ab25b5011 | 22,077 | null | """Wide weight-tied positional loop with numeric digit features and registers.
The pivot
---------
Earlier submissions in this line compressed the prompt into a small global
latent state and iterated an operator on it. That was aimed at the
composition structure of the label, and it did make the composition explicit ... | {
"id": "1d1062b6-5700-4274-9330-3eecdd7031e8",
"created_at": "2026-08-04 13:56:58.757310+00:00",
"db_md5": "bfcd411150d73bf62fcfd1bac9774d1e",
"submitter": "Kaan Karakaya",
"github_login": "TheYkk",
"run_id": "2304596b-6f9c-4b11-bc51-a30fff5a8452",
"tier": "easy",
"dataset_id": "e4",
"status": "succe... | {
"score": {
"mean_loss": 2.4961498207569255,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.006354166679084301
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.4857035290856353,
"example_count": 12... |
1d1878f0-3650-4b67-9a00-57716cf33099 | easy | pepinu | 2026-08-19 06:36:53.820370+00:00 | succeeded | 6b5b6fd71512b1bcb8938e56957a3fc776b5d9442b6c1b823612c049e17bf5e1 | 30,482 | 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": "1d1878f0-3650-4b67-9a00-57716cf33099",
"created_at": "2026-08-19 06:36:53.820370+00:00",
"db_md5": "d82287d6697a27328f0e1b35067c619a",
"submitter": "pepinu",
"github_login": "pepinu",
"run_id": "95cfb4f6-c768-4c03-a825-4674e49487ce",
"tier": "easy",
"dataset_id": "e2",
"status": "succeeded",
... | {
"score": {
"mean_loss": 0.011983036994934082,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 1.0
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 0.01200098916888237,
"example_count": 300,
"... |
1d1ef6bd-41cc-4c6a-90b3-6edc0e5bc72d | easy | sapient-sapiens | 2026-08-14 18:05:05.243183+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": "1d1ef6bd-41cc-4c6a-90b3-6edc0e5bc72d",
"created_at": "2026-08-14 18:05:05.243183+00:00",
"db_md5": "1efadda9e01ae84ec2857432a666c542",
"submitter": "Ertondy",
"github_login": "sapient-sapiens",
"run_id": "664bff14-fbc0-48da-a253-9f774e65ac90",
"tier": "easy",
"dataset_id": "e4",
"status": "su... | {
"score": {
"mean_loss": 2.0673875256446546,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.009479166666666667
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.0706778530701513,
"example_count": 12... |
1d1fc1a3-eef3-4b3d-aeb2-af767dddeafc | easy | LauraGomezjurado | 2026-08-30 23:43:56.923480+00:00 | succeeded | 0e6e3b05c557b6c12d23b25ac2d6e3d2c5107cbf68f7284b2e3f514b60748a91 | 13,398 | null | """Looped depth-recurrent Transformer for repeated modular squaring.
Design notes
------------
The evaluator reads answers off the *last* ``target_len`` prompt positions, so
answers are right-aligned with the least-significant digit at the final valid
position. Digit significance is therefore a fixed offset from the e... | {
"id": "1d1fc1a3-eef3-4b3d-aeb2-af767dddeafc",
"created_at": "2026-08-30 23:43:56.923480+00:00",
"db_md5": "3b5e974075cae097001cc89af5cd642c",
"submitter": "LauraGomezjurado",
"github_login": "LauraGomezjurado",
"run_id": "021e310e-4898-4d59-9d14-35b4db6df0b3",
"tier": "easy",
"dataset_id": "e5",
"st... | {
"score": {
"mean_loss": 2.1778649878269585,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.006250000024835269
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.1767440344720677,
"example_count": 60... |
1d258709-4dc2-4019-bc99-d99c77dc80ff | easy | lzy54 | 2026-08-26 15:04:55.028844+00:00 | failed | 0e91d98b6c8d8db697bca5930d0bd27c7d6b0b21ebf7f84b80519f901477bdfd | 4,504 | null | """Fixed-six-loop Transformer with GPU-masked random effective depth."""
from __future__ import annotations
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
assert_model_state,
)
D_MODEL = 12... | {
"id": "1d258709-4dc2-4019-bc99-d99c77dc80ff",
"created_at": "2026-08-26 15:04:55.028844+00:00",
"db_md5": "88021b7cda1ac447f35a02447cc3f4d1",
"submitter": "lzy54",
"github_login": "lzy54",
"run_id": "8a2ab8e3-b06c-4d07-98cf-dab98088e93b",
"tier": "easy",
"dataset_id": "e6",
"status": "failed",
"st... | null
|
1d291af4-75f3-4770-be1c-c5e19e34872e | easy | nikolageorgiev2000 | 2026-08-04 10:47:10.721585+00:00 | failed | 58e4b5dba1fbaade7c52a19d9d14414aa2aaab7fefdc7c1af14a7af2c09b1f0c | 10,992 | null | """T-composed categorical-digit reasoner with fully learned pair/reduction maps."""
from __future__ import annotations
import math
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
TokenLossBat... | {
"id": "1d291af4-75f3-4770-be1c-c5e19e34872e",
"created_at": "2026-08-04 10:47:10.721585+00:00",
"db_md5": "bd19ab73f5e6b66336c4ad72d993c1d9",
"submitter": "Nikola Georgiev",
"github_login": "nikolageorgiev2000",
"run_id": "e15f199a-cd86-45af-9f32-ad1abbe13add",
"tier": "easy",
"dataset_id": "e2",
"s... | null
|
1d3c6bdc-d413-4022-8248-5f8187722c36 | easy | DDanlov | 2026-08-28 22:05:30.067205+00:00 | succeeded | 79861f2c55de84de896c306403af4bf716741f18d826e9a4dbd716cadb444754 | 15,065 | null | """
Trial 31: Model C: D=512, H=8, M=8, u=2 (T=16)
"""
import math
import torch
import torch.nn as nn
import torch.nn.functional as F
from dataclasses import dataclass
from typing import Optional, Tuple, Dict, Any, List, Union
try:
from benchmark.api import ModelSpec, OptimizerBundle, OptimizerSpec, Submission, T... | {
"id": "1d3c6bdc-d413-4022-8248-5f8187722c36",
"created_at": "2026-08-28 22:05:30.067205+00:00",
"db_md5": "77f9c05f9756eaa6b1ec85bdf04ff9d7",
"submitter": "DDanlov",
"github_login": "DDanlov",
"run_id": "2b60df36-6229-4309-be70-9a06061c99b7",
"tier": "easy",
"dataset_id": "e5",
"status": "succeeded"... | {
"score": {
"mean_loss": 2.2847126032614566,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0141666666790843
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.286309512206792,
"example_count": 600,
... |
1d48a903-2279-4002-bad5-8d00980d31cb | easy | viridale | 2026-08-16 05:35:00.989071+00:00 | succeeded | cb3dfb512a8c0d1a8de2594e398be6cd359b2fbd01afa816c602a7f061e30d8f | 24,505 | null | """Sparse block acquisition over one contextual integer polynomial cell.
hypothesis: joint discrete trust steps recover source-small two-term recurrence edits
from deep final answers even when neither individual coefficient improves loss.
axis: arch
target: invented deep-only r^2+x, r^2-2, and 2r^2-1 controls at T... | {
"id": "1d48a903-2279-4002-bad5-8d00980d31cb",
"created_at": "2026-08-16 05:35:00.989071+00:00",
"db_md5": "e360a813cd89993a28c6c61e5f723b9c",
"submitter": "priormancer",
"github_login": "viridale",
"run_id": "09c14739-b859-4661-a2f2-34829c0242dd",
"tier": "easy",
"dataset_id": "e6",
"status": "succe... | {
"score": {
"mean_loss": 0.15029456466436386,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 1.0
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 0.15949773788452148,
"example_count": 60,
"ex... |
1d55d927-d716-457f-853d-453a77ad6b21 | easy | shirvani-jr | 2026-08-29 04:11:52.587380+00:00 | succeeded | 99bcbe1556e8a8d02da832b06efd2fd65e6659d03c61b60f05ef1f4c7888b310 | 42,893 | null | """E5 research arm: place-shared continuous-workspace recurrent fabric.
The transition is a generic learned local processor. It receives canonical
categorical value and modulus digits, develops a continuous hidden workspace
for a fixed number of internal iterations, then emits the next categorical
value. Workspace n... | {
"id": "1d55d927-d716-457f-853d-453a77ad6b21",
"created_at": "2026-08-29 04:11:52.587380+00:00",
"db_md5": "44f918b47f8a1e4f77eb88e7cf35a400",
"submitter": "Ali",
"github_login": "shirvani-jr",
"run_id": "f84a37e1-2b5e-4f79-bc42-6f27a11aa61a",
"tier": "easy",
"dataset_id": "e5",
"status": "succeeded"... | {
"score": {
"mean_loss": 17.21892270714313,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.010416666896392902
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 17.298206329345703,
"example_count": 600... |
1d5fab59-06cc-441b-ac13-6d3191e44206 | easy | alirezashirvani-jr | 2026-08-28 23:30:57.487504+00:00 | succeeded | 73dad23ac67ce34a8f5ba4257342fbd955f792c46dd910f2668b969b51df72ec | 53,560 | null | """Adaptive learned discrete recurrent fabric for One Layer Deeper.
A gradient-trained, operation-free recurrent computational machine. The
model learns, from the competition's endpoint supervision: how to parse the
prompt into a bank of categorical digit registers, a reusable transition
circuit applied serially, a l... | {
"id": "1d5fab59-06cc-441b-ac13-6d3191e44206",
"created_at": "2026-08-28 23:30:57.487504+00:00",
"db_md5": "c68d20cdc8ff54eb88edee065db13428",
"submitter": "alirezashirvani-jr",
"github_login": "alirezashirvani-jr",
"run_id": "350a00dd-3fc6-4751-b08e-264cd1088d50",
"tier": "easy",
"dataset_id": "e2",
... | {
"score": {
"mean_loss": 2.253219723701477,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.015625000931322575
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.246032476425171,
"example_count": 300,... |
1d611a8e-3792-4e47-8b26-ff331b0d96ac | easy | shirvani-jr | 2026-08-29 02:38:24.989709+00:00 | succeeded | 863b2f964b41c0228e82c26ab7d2f222fdffd3aaa5156fb5da1e115a71ef141f | 48,830 | 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": "1d611a8e-3792-4e47-8b26-ff331b0d96ac",
"created_at": "2026-08-29 02:38:24.989709+00:00",
"db_md5": "8ffc867d4ca7354d9a382379f1c6ff51",
"submitter": "Ali",
"github_login": "shirvani-jr",
"run_id": "b329f57d-f1e3-4d60-af85-a6685e208024",
"tier": "easy",
"dataset_id": "e1",
"status": "succeeded"... | {
"score": {
"mean_loss": 2.0773162841796875,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.04833333194255829
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.0427281856536865,
"example_count": 100... |
1d6157ce-5aea-45a6-b56d-84c0ffbc3c30 | easy | karanganesan | 2026-08-04 01:29:17.944704+00:00 | succeeded | 974ec1eb359c34e313908a0d5394941de7eeb17b1078c99f1a45c9f5374e592e | 11,980 | null | """Parametric looped-transformer family (P1).
One weight-tied transformer block applied k times in latent space. Config
flags cover four P1 families with one file:
looped recall=0 gated=0 tfilm=0 plain weight-tied loop
looped-recall recall=1 re-inject the input embedding each
... | {
"id": "1d6157ce-5aea-45a6-b56d-84c0ffbc3c30",
"created_at": "2026-08-04 01:29:17.944704+00:00",
"db_md5": "bc7ef148a2426f1c6c90e28e9e05314c",
"submitter": "Karan Ganesan",
"github_login": "karanganesan",
"run_id": "96b723dd-caae-4adf-9d4d-235af93a3297",
"tier": "easy",
"dataset_id": "e5",
"status": ... | {
"score": {
"mean_loss": 2.719480514526367,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.006666666828095913
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.74753999710083,
"example_count": 600,
... |
1d6ae5bf-e1e7-4f53-a8bc-372d1f821b5f | easy | repst | 2026-08-29 22:54:02.973860+00:00 | succeeded | 26ba375520b4b1f08562b1ee0ffb6f378758b74377c0f8f8171aed8d709fafde | 22,654 | null | """R876: replay-trained recurrent Markov Transformer.
Authored structure, analogous to an LM's tokenizer and neural architecture:
* marker-delimited decimal N/X/T fields and right-aligned digit places;
* a learned decimal vocabulary shared by input, recurrent state, and output;
* a near-identity, T-blind neural input... | {
"id": "1d6ae5bf-e1e7-4f53-a8bc-372d1f821b5f",
"created_at": "2026-08-29 22:54:02.973860+00:00",
"db_md5": "023ceb7905a54c6f88d20ac2a3a0cfaf",
"submitter": "Ryan Epstein",
"github_login": "repst",
"run_id": "032c751d-3890-44f8-b5da-e9d7c3da2050",
"tier": "easy",
"dataset_id": "e1",
"status": "succeed... | {
"score": {
"mean_loss": 1.2996866106987,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.6316666603088379
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 0.9856612682342529,
"example_count": 100,
... |
1d7826b1-844a-4785-8471-d3a5b343015c | easy | ehonig | 2026-08-14 06:05:01.631567+00:00 | succeeded | 00eeaf71e1ef7eb8bca1dce42d5b5f824a664f5246edf93d90ecb8ba73c377e0 | 52,555 | null | """DigitLoop — a place-value recurrent model for repeated modular squaring.
The task is ``y = x^(2^T) mod N`` with the answer read back tail aligned, one
decimal place value per token position from the right. A stack of attention and
MLP layers has no primitive for the two operations this needs — a partial-product
co... | {
"id": "1d7826b1-844a-4785-8471-d3a5b343015c",
"created_at": "2026-08-14 06:05:01.631567+00:00",
"db_md5": "5e27483be6be1e0b65bea8bf17fdfe9a",
"submitter": "Edouardo Honig",
"github_login": "ehonig",
"run_id": "9363916f-fd97-4b4e-8a35-0f12e40f3401",
"tier": "easy",
"dataset_id": "e1",
"status": "succ... | {
"score": {
"mean_loss": 4.010694146156311,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.4099999964237213
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 3.1688711643218994,
"example_count": 100,
... |
1d79d0c1-1df4-40f1-81c8-df5483e659b0 | easy | repst | 2026-08-05 20:25:46.338444+00:00 | succeeded | 00c4dd83c7d4acd1b25894ea6572181318a550567ba656dad9fa5ee4263540cf | 6,248 | null | """Small tied-recurrent Transformer with intermediate supervision."""
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,... | {
"id": "1d79d0c1-1df4-40f1-81c8-df5483e659b0",
"created_at": "2026-08-05 20:25:46.338444+00:00",
"db_md5": "a542a4bdd7e0ac86f98578d0c1109bcb",
"submitter": "Ryan Epstein",
"github_login": "repst",
"run_id": "54814018-c783-479b-93dc-9627eb746c44",
"tier": "easy",
"dataset_id": "e1",
"status": "succeed... | {
"score": {
"mean_loss": 2.092949628829956,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.07333333417773247
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.1265408992767334,
"example_count": 100,... |
1d7ddd56-053b-4ea7-9439-36f753d78e8e | easy | liam-gb | 2026-08-11 01:41:43.129541+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": "1d7ddd56-053b-4ea7-9439-36f753d78e8e",
"created_at": "2026-08-11 01:41:43.129541+00:00",
"db_md5": "eaac69fe345f82973fcf392c504a40ff",
"submitter": "liam-gb",
"github_login": "liam-gb",
"run_id": "5ecb98a0-4a39-4d1a-9958-2898f990415e",
"tier": "easy",
"dataset_id": "e1",
"status": "succeeded"... | {
"score": {
"mean_loss": 2.858112096786499,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.014999999664723873
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 3.7183303833007812,
"example_count": 100... |
1d88f9a1-938a-4832-97e6-44e74f1979ba | easy | yashkant | 2026-08-15 14:30:33.447158+00:00 | succeeded | 882409b73583024c5f534f37a403a2301c0a145ddd2b2bafb06ac29bf63f6e2a | 20,745 | null | """Round1956 Round493 block-two shared-head supervision.
The exact Round493 model, initialization, optimizer, and final prediction path
remain intact. During training both arms additionally read out the hidden state
after encoder block two with the already-existing norm, attention pool,
decoder, and answer head. The c... | {
"id": "1d88f9a1-938a-4832-97e6-44e74f1979ba",
"created_at": "2026-08-15 14:30:33.447158+00:00",
"db_md5": "44bd6c90bb314307b60ffde74f35f1ac",
"submitter": "Yash Kant",
"github_login": "yashkant",
"run_id": "676062f4-4932-4340-b521-d36b26e3eeb7",
"tier": "easy",
"dataset_id": "e4",
"status": "succeed... | {
"score": {
"mean_loss": 2.0624593508920146,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0077083333333333335
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.064892513453812,
"example_count": 12... |
1d908aa9-8c1e-43b6-b184-e0ecde7cf054 | easy | nikolageorgiev2000 | 2026-08-04 00:35:24.241597+00:00 | succeeded | 42eb187ee1fcc577699f5bf503172347336ea31820fb07f7952f1cc903dd2ef7 | 8,394 | null | """K=1 recurrent graph reasoner with explicit differentiable writable memory."""
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_m... | {
"id": "1d908aa9-8c1e-43b6-b184-e0ecde7cf054",
"created_at": "2026-08-04 00:35:24.241597+00:00",
"db_md5": "1cb6961aed2e74d38b33161eab2e4ed7",
"submitter": "Nikola Georgiev",
"github_login": "nikolageorgiev2000",
"run_id": "e91127d8-8e4a-42c6-8923-b3ed817c31af",
"tier": "easy",
"dataset_id": "e1",
"s... | {
"score": {
"mean_loss": 2.4318509101867676,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.04666666593402624
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.7495148181915283,
"example_count": 100... |
1d9d3b73-f949-4cf8-b3c2-71f43babb6a7 | easy | khushidahi | 2026-08-02 06:22:00.918890+00:00 | succeeded | 4446fb87dceaabf2a4d0de670ff6d1eb3e51dfda6dc04c5f6161ff12960b7b36 | 8,484 | 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": "1d9d3b73-f949-4cf8-b3c2-71f43babb6a7",
"created_at": "2026-08-02 06:22:00.918890+00:00",
"db_md5": "7ad94e3e9d48f054bbb9940ee51e8acf",
"submitter": "khushidahi",
"github_login": "khushidahi",
"run_id": "26f01779-f624-4ecf-b025-666ac36625d6",
"tier": "easy",
"dataset_id": "e5",
"status": "succ... | {
"score": {
"mean_loss": 2.1679468154907227,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.009166666772216558
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.1754045486450195,
"example_count": 60... |
1d9ef28f-461c-4500-81d7-71faec1dc36e | easy | DDanlov | 2026-08-14 07:58:37.396806+00:00 | succeeded | 424ecfd803cae7f86f3798610fa5a76489a75adf925dfea4ef04181e034fcd26 | 21,975 | null | import math
from typing import Optional
import torch
import torch.nn as nn
import torch.nn.functional as F
from torch.optim import Optimizer
try:
from benchmark import Submission, assert_model_state, OptimizerBundle
except ImportError:
try:
from client import Submission, assert_model_state, OptimizerBu... | {
"id": "1d9ef28f-461c-4500-81d7-71faec1dc36e",
"created_at": "2026-08-14 07:58:37.396806+00:00",
"db_md5": "08aec454fc7a642d370cb9f3a3ce4ef7",
"submitter": "DDanlov",
"github_login": "DDanlov",
"run_id": "c469c2e2-706e-4b7a-8106-ccb1841a5324",
"tier": "easy",
"dataset_id": "e5",
"status": "succeeded"... | {
"score": {
"mean_loss": 2.3105110754317435,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0141666666790843
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.3138626955549815,
"example_count": 600,... |
1da3ba2b-35f5-4ce7-993d-5a076291a3d5 | easy | shreyash-chonkie | 2026-09-01 03:18:48.889318+00:00 | succeeded | bd16c29015b34ae68cf817fec2b39e756da156ed097ff67633592190eb5f2a78 | 20,008 | null | """Learned token scanner feeding a random recurrent path reservoir."""
from __future__ import annotations
from itertools import permutations
import torch
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
TokenLossBatch,
assert_model_s... | {
"id": "1da3ba2b-35f5-4ce7-993d-5a076291a3d5",
"created_at": "2026-09-01 03:18:48.889318+00:00",
"db_md5": "c52a7411490c902802878dfb66f3fe6b",
"submitter": "Shreyash",
"github_login": "shreyash-chonkie",
"run_id": "d708b3e1-7139-452c-a270-2e2923bb0ab2",
"tier": "easy",
"dataset_id": "e2",
"status": "... | {
"score": {
"mean_loss": 24.893590927124023,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.026666668942198157
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 25.86597442626953,
"example_count": 300... |
1da4db34-a886-4f1c-9a41-4b98149c7993 | easy | khushidahi | 2026-08-05 20:29:15.581036+00:00 | succeeded | b22de50629766cb101d9d27251a96f63d765768fc3db076316bd6d45929e1e55 | 15,206 | 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": "1da4db34-a886-4f1c-9a41-4b98149c7993",
"created_at": "2026-08-05 20:29:15.581036+00:00",
"db_md5": "57a3e9fe2638508e886042d978d32eff",
"submitter": "khushidahi",
"github_login": "khushidahi",
"run_id": "65c83cb7-6cf7-4cc2-848a-57c162d5c221",
"tier": "easy",
"dataset_id": "e5",
"status": "succ... | {
"score": {
"mean_loss": 2.2081714868545532,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.009583333507180214
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.204859733581543,
"example_count": 600... |
1dab6a71-752f-4b71-924e-756d761a92b4 | easy | YuMingQian1234 | 2026-08-07 05:44:42.481401+00:00 | succeeded | cdd81c773e622804e9464851091d4d87844e243287ffa91c8e43eb52e31f3e30 | 7,193 | null | """Simplified LoopFormer for one-layer-deeper submission."""
from __future__ import annotations
import math
from dataclasses import dataclass
import torch
import torch.nn as nn
import torch.nn.functional as F
from torch import Tensor
from benchmark import ModelSpec, OptimizerBundle, OptimizerSpec, Submission, asser... | {
"id": "1dab6a71-752f-4b71-924e-756d761a92b4",
"created_at": "2026-08-07 05:44:42.481401+00:00",
"db_md5": "6886c92ac60b5153e33ed3f1f5e1246e",
"submitter": "Mingqian Yu",
"github_login": "YuMingQian1234",
"run_id": "6ab01c2c-8905-4ff8-974d-00eb3f5129ee",
"tier": "easy",
"dataset_id": "e1",
"status": ... | {
"score": {
"mean_loss": 2.0319408178329468,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.054999999701976776
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.0212442874908447,
"example_count": 10... |
1dafe071-a649-4451-89eb-0c99e6e2386e | easy | khushidahi | 2026-08-06 02:54:09.781045+00:00 | succeeded | abb737a4c18cc545ab64e1ad0e06a24a02f03bf8a0cca6b8ef722855c17071be | 14,389 | 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": "1dafe071-a649-4451-89eb-0c99e6e2386e",
"created_at": "2026-08-06 02:54:09.781045+00:00",
"db_md5": "0ff93c1a4c51b850d85208408dfb3fe5",
"submitter": "khushidahi",
"github_login": "khushidahi",
"run_id": "e5387939-647a-4fa3-bcad-d2e2de96d67f",
"tier": "easy",
"dataset_id": "e2",
"status": "succ... | {
"score": {
"mean_loss": 2.23936665058136,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.009583333972841501
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.237718343734741,
"example_count": 300,
... |
1db347de-88a5-4b44-884d-1e89be3e72d7 | easy | DDanlov | 2026-08-13 22:43:35.598894+00:00 | failed | 9510906704a0119f36cd9b72253ea4d072576200e7dbf2681a696a121b193d3a | 22,142 | null | import torch
import torch.nn as nn
import torch.nn.functional as F
import math
from dataclasses import dataclass
from typing import Optional
from torch.optim import Optimizer
try:
from client import Submission, assert_model_state, OptimizerBundle
except ImportError:
Submission = None
assert_model_state = N... | {
"id": "1db347de-88a5-4b44-884d-1e89be3e72d7",
"created_at": "2026-08-13 22:43:35.598894+00:00",
"db_md5": "91328cd7240c3741d722f83fc5485937",
"submitter": "DDanlov",
"github_login": "DDanlov",
"run_id": "68eb509a-eba1-493f-a2e3-924fa8bb52c7",
"tier": "easy",
"dataset_id": "e3",
"status": "failed",
... | null
|
1db66bab-2292-4212-b5c7-1c46ee4d52bb | easy | himalalps | 2026-08-06 08:48:57.301225+00:00 | succeeded | 93456741d08e5535f1405f907a6153978255f09211cf8ce8cb459bc2b3c9a4ea | 27,775 | null | """Algorithmically constrained latent recurrence for One Layer Deeper.
One masked cross-attention pass creates two independent representations:
``state_0`` from X only and an immutable context from N only. T is excluded
from both representations and is used solely to choose how many times the
learned recurrent MLP tr... | {
"id": "1db66bab-2292-4212-b5c7-1c46ee4d52bb",
"created_at": "2026-08-06 08:48:57.301225+00:00",
"db_md5": "388b781e6eb7011249ac3719c8da9d8f",
"submitter": "Haoyu Tang",
"github_login": "himalalps",
"run_id": "c2c5f87f-9653-4cec-8f84-c711eb67b2f7",
"tier": "easy",
"dataset_id": "e3",
"status": "succe... | {
"score": {
"mean_loss": 15.45585501811633,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.02312500015832484
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 15.953055424704289,
"example_count": 800,... |
1db78288-eaea-4daa-b0c2-85aed338905d | easy | alirezashirvani-jr | 2026-08-30 22:00:59.274596+00:00 | succeeded | d90e616280ac0e2842d80d3c9dc7a3a5397a8286ca90e919b3e4aa4e0a6ca12b | 25,012 | null | from __future__ import annotations
import math
import time
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import ModelSpec, OptimizerBundle, OptimizerSpec, Submission, TokenLossBatch, assert_model_state
EXPERT_COUNT = 12
PAIR_COUNT = EXPERT_COUNT * (EXPERT_COUNT - 1) // 2
... | {
"id": "1db78288-eaea-4daa-b0c2-85aed338905d",
"created_at": "2026-08-30 22:00:59.274596+00:00",
"db_md5": "6efdb492b3a449315ab62e78ee3408e9",
"submitter": "alirezashirvani-jr",
"github_login": "alirezashirvani-jr",
"run_id": "735d9fa4-fc89-4faf-abdd-9f6c6423bd49",
"tier": "easy",
"dataset_id": "e3",
... | {
"score": {
"mean_loss": 5.801888704299927,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.006249999976716936
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 5.627822399139404,
"example_count": 800,... |
1dbec7ab-c625-408b-94d4-e585bb663236 | easy | DDanlov | 2026-08-22 23:34:53.697301+00:00 | succeeded | 124b8e5748a370d39273d61e757fa3549f243c3d0569add1965f147c1c5a263c | 27,587 | null | """
Unified Hybrid Masked Deliberation Model (75.8M Parameters) - Variant 2 (Pure Skipless, No Cross-Attention)
Architecture:
- Scale: dim=1536, num_heads=24, d_ff=6144 (Exactly 75,826,176 parameters)
- Hybrid Prefix-Bidirectional + Causal Target Attention Mask
- Token Layout: [Prompt (L_p), 4 Pause Registers, Target T... | {
"id": "1dbec7ab-c625-408b-94d4-e585bb663236",
"created_at": "2026-08-22 23:34:53.697301+00:00",
"db_md5": "c77213c90ca2f8b17d7c294303018991",
"submitter": "DDanlov",
"github_login": "DDanlov",
"run_id": "65aea7bc-3fda-426d-a670-264c7bbfdab1",
"tier": "easy",
"dataset_id": "e5",
"status": "succeeded"... | {
"score": {
"mean_loss": 22.228872092958362,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0004166666666666667
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 22.197847143951552,
"example_count": 6... |
1dc7293a-3395-41e0-9032-12e506956dbf | easy | shirvani-jr | 2026-08-28 17:13:05.693704+00:00 | succeeded | 3a21df34102420cf6fd6e416204ec266aedbf31732c3ead62864db06fa9f92be | 53,900 | 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": "1dc7293a-3395-41e0-9032-12e506956dbf",
"created_at": "2026-08-28 17:13:05.693704+00:00",
"db_md5": "23ba26ea575999d374bf33e9513f8bdf",
"submitter": "Ali",
"github_login": "shirvani-jr",
"run_id": "a29554f7-e7e5-49f9-87f9-c48f93c31dd8",
"tier": "easy",
"dataset_id": "e1",
"status": "succeeded"... | {
"score": {
"mean_loss": 1.9255062341690063,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.06000000052154064
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 1.8818788528442383,
"example_count": 100... |
1dc9b8fa-f3c6-46f0-8947-0057453efe1b | easy | lpbb | 2026-08-14 10:29:51.705413+00:00 | succeeded | aab5329d5cfc081bac459cc14edb137b2f417dd206323856334b71f63cffbec2 | 8,017 | null | """CHEAT L0b — the hint IS the answer; the network only has to learn to copy the
4 hint digits to the 4 answer slots. Measures the pure transport/routing ceiling
so L1's score decomposes into transport vs arithmetic."""
from __future__ import annotations
import math
import torch
import torch.nn.functional as F
from ... | {
"id": "1dc9b8fa-f3c6-46f0-8947-0057453efe1b",
"created_at": "2026-08-14 10:29:51.705413+00:00",
"db_md5": "69ff206b41ebb547791c6040e2c445c4",
"submitter": "Lpbb",
"github_login": "lpbb",
"run_id": "2957fa5a-b10a-45c0-8613-88256a595ec8",
"tier": "easy",
"dataset_id": "e3",
"status": "succeeded",
"s... | {
"score": {
"mean_loss": 0.1925922632473139,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.9287499988079071
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 0.3749980764086627,
"example_count": 800,... |
1dd43c76-444f-453e-9839-3b7679de1fb0 | easy | sapient-sapiens | 2026-08-23 08:04:02.387770+00:00 | succeeded | a1afb9fde8cca806267e38746a0248580c9f50fe22470e21e9917c48ff8ab560 | 19,748 | null | """Approximately 10x-scale adaptive child of the D6 pair/quotient architecture.
The architecture uses the required learned additive token/group/left-relative
representation. Marker-derived field lengths additionally route categorical
digit embeddings into right-aligned lanes; this alternative place alignment is
an ex... | {
"id": "1dd43c76-444f-453e-9839-3b7679de1fb0",
"created_at": "2026-08-23 08:04:02.387770+00:00",
"db_md5": "19bda25ab7c1be60edcefb46de545a93",
"submitter": "Ertondy",
"github_login": "sapient-sapiens",
"run_id": "b028b6cb-ee51-45b4-bb9f-24abc8911406",
"tier": "easy",
"dataset_id": "e5",
"status": "su... | {
"score": {
"mean_loss": 2.1970216395350795,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0008333333333333334
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.198239099952672,
"example_count": 60... |
1dd45c66-afb6-456a-a12f-80329b9ff11d | easy | khushidahi | 2026-08-02 05:09:31.860137+00:00 | succeeded | 209e9f355db9fc351748182bb717d27676badd4524dbe14c35076489d86cc4b2 | 8,464 | 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": "1dd45c66-afb6-456a-a12f-80329b9ff11d",
"created_at": "2026-08-02 05:09:31.860137+00:00",
"db_md5": "0b73ebd15390e0e338a5c1afb95f36a1",
"submitter": "khushidahi",
"github_login": "khushidahi",
"run_id": "51053425-291a-4844-b2bf-6b52e7f2d0fc",
"tier": "easy",
"dataset_id": "e5",
"status": "succ... | {
"score": {
"mean_loss": 2.1732547283172607,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.009166667005047202
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.1807100772857666,
"example_count": 60... |
1dd73903-1940-4aa2-8c3d-3367b64270de | easy | 0Chris5R | 2026-08-07 09:03:31.795858+00:00 | succeeded | f5ef15c83712f311ecc3d245b69f4f81dd7d769cea7cf303ef8518bb8b921e88 | 30,692 | 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": "1dd73903-1940-4aa2-8c3d-3367b64270de",
"created_at": "2026-08-07 09:03:31.795858+00:00",
"db_md5": "9e14d50e825c1ee500eae83dad30f9e3",
"submitter": "Chris ",
"github_login": "0Chris5R",
"run_id": "ca6c761d-9a3d-4826-8f94-8823c2b18e25",
"tier": "easy",
"dataset_id": "e5",
"status": "succeeded"... | {
"score": {
"mean_loss": 2.2971170188359675,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.008750000012417634
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.2089444479065627,
"example_count": 60... |
1ddb6596-ed83-42cc-aff1-84ad666756f7 | easy | k-penchev | 2026-08-22 23:46:03.973021+00:00 | succeeded | 7f3888911ec44edc0642940193eb2ec64b2992b3330cc4c65b54326a4437a979 | 8,850 | null | """Schedule-Free gated recurrence with Adam-style inner momentum."""
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,
as... | {
"id": "1ddb6596-ed83-42cc-aff1-84ad666756f7",
"created_at": "2026-08-22 23:46:03.973021+00:00",
"db_md5": "1ddf54122c2ec060b40e3a48709b54c6",
"submitter": "Kaloyan Penchev",
"github_login": "k-penchev",
"run_id": "c7f4ac8b-b4c8-433e-b8ff-3a1c6ba84699",
"tier": "easy",
"dataset_id": "e1",
"status": "... | {
"score": {
"mean_loss": 1.955591857433319,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.018333333544433117
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 1.9370697736740112,
"example_count": 100... |
1ddfd5e0-9a89-4fad-907f-152f9399d397 | easy | MarcoF1 | 2026-08-13 20:34:53.322540+00:00 | succeeded | e0818ed32bd90197c94b0af1a411d1915758a6e78c6cd7d42a5898149844eb76 | 14,397 | null | """Residue-clock model for fixed-N generalization.
Representation: parsed integer values of x and N feed (a) exact residue-phase
features over a bank of small integer periods, (b) learned double-hashed value
embeddings (per-residue identity), and T feeds cyclic phase features
(T mod 2..48) plus saturating channels — p... | {
"id": "1ddfd5e0-9a89-4fad-907f-152f9399d397",
"created_at": "2026-08-13 20:34:53.322540+00:00",
"db_md5": "56f4b8f6d3ca7b791ac93432c3184691",
"submitter": "Marco Fleming",
"github_login": "MarcoF1",
"run_id": "f837c854-21fc-46e0-82ce-9ac99298e7e6",
"tier": "easy",
"dataset_id": "e4",
"status": "succ... | {
"score": {
"mean_loss": 2.106178609270967,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0032291666666666666
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.119771444684662,
"example_count": 120... |
1df10205-8b94-405f-8320-2a1aad2d66b5 | easy | himalalps | 2026-08-09 12:36:21.102067+00:00 | succeeded | 2d99bb2f4b37794dcf1c39d0178347d644d81d9afa7089ea191c4020a7fa1027 | 81,240 | null | """Algorithmically constrained latent recurrence for One Layer Deeper.
One masked cross-attention pass creates two independent representations:
``state_0`` from X only and an immutable context from N only. T is excluded
from both representations and is used solely to choose how many times the
learned recurrent MLP tr... | {
"id": "1df10205-8b94-405f-8320-2a1aad2d66b5",
"created_at": "2026-08-09 12:36:21.102067+00:00",
"db_md5": "d938ba697903aff13e8e49aa1ee0f128",
"submitter": "Haoyu Tang",
"github_login": "himalalps",
"run_id": "866b193b-e227-4148-a867-27fcbc74e61e",
"tier": "easy",
"dataset_id": "e5",
"status": "succe... | {
"score": {
"mean_loss": 2.181654635682568,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.010416666896392902
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.165496349334717,
"example_count": 600,... |
1df4d02a-3b83-4cde-b679-c70af14ab5e3 | easy | DDanlov | 2026-08-11 23:01:13.095583+00:00 | succeeded | 4a282732852ddc83a29e956404dbe80e8758fbab59fd28f73a9d0c29eabd13cb | 21,864 | 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": "1df4d02a-3b83-4cde-b679-c70af14ab5e3",
"created_at": "2026-08-11 23:01:13.095583+00:00",
"db_md5": "5ed9774dac19db5dc4268f71abc81fff",
"submitter": "DDanlov",
"github_login": "DDanlov",
"run_id": "97bc86a5-091c-4893-8b17-5215107272f8",
"tier": "easy",
"dataset_id": "e5",
"status": "succeeded"... | {
"score": {
"mean_loss": 75480.94548304731,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.005
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 76460.12496058717,
"example_count": 600,
"exa... |
1dfbb8ba-4c08-4507-84ad-e6c60b5d3951 | easy | AFeinfield | 2026-08-28 22:33:14.037837+00:00 | succeeded | d30cab20b0ed18787abda86f6d89b03913cff886154e5a1449a162c39e0c51d2 | 14,291 | null | """195: drilled machine, throughput/convergence pass: cumprod compare fold
(2 vectorized ops vs a 16-iteration loop) and table-param lr 0.2 so
drills snap in tens of steps.
All six operator tables (mul lo/hi, add+carry, sub+borrow, compare) are
learned classifications over digit pairs, supervised by GENERIC base-10
dr... | {
"id": "1dfbb8ba-4c08-4507-84ad-e6c60b5d3951",
"created_at": "2026-08-28 22:33:14.037837+00:00",
"db_md5": "e3773c1f2cddcd79a234b2aac0af28f5",
"submitter": "AFeinfield",
"github_login": "AFeinfield",
"run_id": "2d8888df-820a-4505-a1ea-949a9f4c3ba0",
"tier": "easy",
"dataset_id": "e1",
"status": "succ... | {
"score": {
"mean_loss": 2.8287293910980225,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.8286056518554688,
"example_count": 100,
"exa... |
1e05f611-d43e-4a8c-b6ca-0920ed3eeda5 | easy | himalalps | 2026-08-09 03:36:33.111593+00:00 | succeeded | 090365a2ae1d57b2bbb262c21629dd6c85d4a7a112e5fff7c3c6a0215d793f83 | 73,648 | null | """Algorithmically constrained latent recurrence for One Layer Deeper.
One masked cross-attention pass creates two independent representations:
``state_0`` from X only and an immutable context from N only. T is excluded
from both representations and is used solely to choose how many times the
learned recurrent MLP tr... | {
"id": "1e05f611-d43e-4a8c-b6ca-0920ed3eeda5",
"created_at": "2026-08-09 03:36:33.111593+00:00",
"db_md5": "a95b6e50ea41205a94cda6127def7259",
"submitter": "Haoyu Tang",
"github_login": "himalalps",
"run_id": "c74b6ec4-feb2-4ae5-9862-4de826935a12",
"tier": "easy",
"dataset_id": "e1",
"status": "succe... | {
"score": {
"mean_loss": 1.897165834903717,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.07333333045244217
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 1.7850254774093628,
"example_count": 100,... |
1e07c80e-e42d-4dc2-8140-d97f5cec5ca1 | easy | yunjiangster | 2026-08-27 21:22:58.771969+00:00 | succeeded | cab18f706f9fb725ac758e730cb6cb23c13666991416f7de91b4128756ced19a | 14,059 | null | """Generic learned-program probe for One Layer Deeper.
A randomly initialized categorical parameter selects one program from a generic
accumulator-machine DSL. Programs are generated mechanically from four generic
instructions over the public scalar input fields; no modular-squaring expression
or arithmetic trace is ... | {
"id": "1e07c80e-e42d-4dc2-8140-d97f5cec5ca1",
"created_at": "2026-08-27 21:22:58.771969+00:00",
"db_md5": "0baadc0045903028c9b11d7d3c552c5c",
"submitter": "Yunjiang Jiang",
"github_login": "yunjiangster",
"run_id": "6a15bac4-f37a-464a-b53c-ae1acf22aedb",
"tier": "easy",
"dataset_id": "e5",
"status":... | {
"score": {
"mean_loss": 0.004424781072884798,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 1.0
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 0.004424781072884798,
"example_count": 600,
... |
1e08105a-367f-4f8d-ab1a-d199dfaf9c78 | easy | sapient-sapiens | 2026-08-13 20:29:17.243318+00:00 | succeeded | 829f0a7ed78146134ceca61a39e343b7b2961c37feb5ac85e6884b98612c34dc | 6,031 | null | """Paper-style additive-input-injection loop for fixed-T=2 E3."""
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, assert_model_state
WIDTH = 1536
HEADS =... | {
"id": "1e08105a-367f-4f8d-ab1a-d199dfaf9c78",
"created_at": "2026-08-13 20:29:17.243318+00:00",
"db_md5": "2f3f403cd6dccd63981f30b69cf1eac3",
"submitter": "Ertondy",
"github_login": "sapient-sapiens",
"run_id": "680b0ea3-62c4-4ec3-aa6e-1849a37b60ec",
"tier": "easy",
"dataset_id": "e3",
"status": "su... | {
"score": {
"mean_loss": 2.220091860114347,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.01
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.2154673904142914,
"example_count": 800,
"exa... |
1e09a7f9-513d-43d0-9504-3eccf3705cc7 | easy | viridale | 2026-08-02 21:06:04.924799+00:00 | succeeded | fc2ce47faf4f935fa4df29dcf99c057c0dff2d4100b1647b30f585e444018481 | 67,145 | null | """Minimum-depth sequential expansion with learned staged carry.
hypothesis: v113 bounds family-8 depth but redundantly rescans all rejected
lower catalogues on updates2/3. Carry the learned top512 from each stage,
score those active complete rules on the current rows, and exhaustively
score only the newl... | {
"id": "1e09a7f9-513d-43d0-9504-3eccf3705cc7",
"created_at": "2026-08-02 21:06:04.924799+00:00",
"db_md5": "3548734bfdb46acf325b1f845619a11a",
"submitter": "priormancer",
"github_login": "viridale",
"run_id": "bcb4c7a5-0097-43db-bbe6-d8ff170c7318",
"tier": "easy",
"dataset_id": "e1",
"status": "succe... | {
"score": {
"mean_loss": 0.003014546020381487,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 1.0
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 0.0030145461205393076,
"example_count": 100,
... |
1e11bba4-7438-47f8-8118-d65721c57420 | easy | LauraGomezjurado | 2026-08-30 21:30:59.916027+00:00 | succeeded | 38848aec07dd988276e529adbf3e2adf3c6d985cc356fddb21ba07920205fe4b | 9,045 | null | """Looped depth-recurrent Transformer for repeated modular squaring.
Design notes
------------
The evaluator reads answers off the *last* ``target_len`` prompt positions, so
answers are right-aligned with the least-significant digit at the final valid
position. Digit significance is therefore a fixed offset from the e... | {
"id": "1e11bba4-7438-47f8-8118-d65721c57420",
"created_at": "2026-08-30 21:30:59.916027+00:00",
"db_md5": "a900a4d76f8195b9f1e6aa0566a23d2f",
"submitter": "LauraGomezjurado",
"github_login": "LauraGomezjurado",
"run_id": "9df981ab-7fd7-48f8-a2b7-3f2248c0ead4",
"tier": "easy",
"dataset_id": "e7",
"st... | {
"score": {
"mean_loss": 1.9090188807911344,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.048752229434974266
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 1.8522541522979736,
"example_count": 85... |
1e145163-4301-4ab4-a324-1d848841afd2 | easy | AFeinfield | 2026-08-27 22:10:18.385069+00:00 | succeeded | 63658156ae1a55f35497e4d0bf2d244d792a63cb7993e52f89073f72d53760bc | 11,982 | null | """151: 149 + throughput/learnability pass for the easy-tier loop.
Dynamic alignments (m <= batch N-width), carry passes 6 post-square / 3
in reduction, single correction try, explicit y/10 feature (carry is
linear in it), quotient MLP 48->96, lr 1e-2.
149 notes: LEARNED QUOTIENT ESTIMATION.
148 learned (loss 2.85->... | {
"id": "1e145163-4301-4ab4-a324-1d848841afd2",
"created_at": "2026-08-27 22:10:18.385069+00:00",
"db_md5": "04cbe5856313ce97b1b1b0763c4587f1",
"submitter": "AFeinfield",
"github_login": "AFeinfield",
"run_id": "2807d239-0fb9-45e4-a3c7-73b56b2719b2",
"tier": "easy",
"dataset_id": "e5",
"status": "succ... | {
"score": {
"mean_loss": 2.1713193026891835,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0037500000124176345
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.175164231270418,
"example_count": 60... |
1e153190-26a7-4a4d-91d3-083f58bbb278 | easy | rteehas | 2026-08-19 11:54:39.613144+00:00 | succeeded | 530b06f816f7bb15d366b8efbc541bfa84d38a4f20fa8ed46f7a588a2fe71e65 | 56,207 | null | """Anonymous marginal-program submission.
Mechanism
---------
The model is a differentiable probabilistic program. Its forward pass
computes, exactly, the answer-token predictions of a small structured latent
hypothesis space:
(prompt segmentation = which token types head the three fields)
x (field-role permutat... | {
"id": "1e153190-26a7-4a4d-91d3-083f58bbb278",
"created_at": "2026-08-19 11:54:39.613144+00:00",
"db_md5": "4fac5f769f8d97bb41ffd3368b1f4a51",
"submitter": "Ryan Teehan",
"github_login": "rteehas",
"run_id": "7acd78e6-4574-4059-903d-a6d43ba7ad80",
"tier": "easy",
"dataset_id": "e2",
"status": "succee... | {
"score": {
"mean_loss": 2.833213448524475,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.8332133293151855,
"example_count": 300,
"exac... |
1e1c0aba-70cf-4cda-bbe0-55831f427f56 | easy | oupadhyay | 2026-08-10 20:09:04.769491+00:00 | succeeded | 99e5ce3680a6f630890399df85e5bc8a8750b094b36cedb763aedf43bf0af406 | 9,944 | null | """Fixed eight-tick differentiable neural-stack transition for FNST-8 Easy E5."""
from __future__ import annotations
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import ModelSpec, OptimizerBundle, OptimizerSpec, Submission, assert_model_state
PAD, N_MARK, X_MARK, T_MARK, A... | {
"id": "1e1c0aba-70cf-4cda-bbe0-55831f427f56",
"created_at": "2026-08-10 20:09:04.769491+00:00",
"db_md5": "eaec154226ad888ec9dbfa847bdf45f5",
"submitter": "Ojasw Upadhyay",
"github_login": "oupadhyay",
"run_id": "b3972b39-482f-4048-ad92-75a52b8f3622",
"tier": "easy",
"dataset_id": "e5",
"status": "s... | {
"score": {
"mean_loss": 2.816596513228707,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.003750000024835269
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.827794253558856,
"example_count": 600,... |
1e1d4ab1-9856-4c29-805d-94881f80cbfd | easy | Yalyenea | 2026-08-11 20:55:02.031140+00:00 | succeeded | 7128b5839d2b01f7c32b3805b64056786c7a9092a475026777876834610e3ee7 | 12,161 | null | """R4 parameter-matched SwiGLU full-bandwidth candidate with cosine horizon 2000."""
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_mode... | {
"id": "1e1d4ab1-9856-4c29-805d-94881f80cbfd",
"created_at": "2026-08-11 20:55:02.031140+00:00",
"db_md5": "113ca773f05e0dd54458d078650b9730",
"submitter": "yfff",
"github_login": "Yalyenea",
"run_id": "028e498e-32d0-454d-9331-70ca9fc5fd33",
"tier": "easy",
"dataset_id": "e5",
"status": "succeeded",
... | {
"score": {
"mean_loss": 7.026496020254788,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.009583333414047957
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 6.8646321296691895,
"example_count": 600... |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.