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 |
|---|---|---|---|---|---|---|---|---|---|---|
2f165a7f-a1a8-4867-b350-13062595b4d0 | easy | khushidahi | 2026-08-03 22:22:54.750572+00:00 | succeeded | d7797656025009174779ba680cbbb40b4f64bf5f8a953be639300ef590507548 | 15,555 | 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": "2f165a7f-a1a8-4867-b350-13062595b4d0",
"created_at": "2026-08-03 22:22:54.750572+00:00",
"db_md5": "2d746f6a3ef1f5cd9a13e51f82744c29",
"submitter": "khushidahi",
"github_login": "khushidahi",
"run_id": "94294677-f156-481c-832b-a936ad3a68b0",
"tier": "easy",
"dataset_id": "e5",
"status": "succ... | {
"score": {
"mean_loss": 2.197807192802429,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.008333333302289248
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.1933329105377197,
"example_count": 600... |
2f20994f-f203-4cb0-b3dc-52993e1cbdd5 | easy | yashkant | 2026-08-10 20:25:47.501642+00:00 | succeeded | 8e3ad45dde3bc081bffc8263034b1e05b611d49e69651cefad2f7bc53a779744 | 31,137 | null | """Round1282 shared continued-fraction denominator-recursion pair.
The complete Round1141 D416 MemGLU carrier remains byte-identical. Its pooled
representation seeds a positive terminal denominator. Every shared call adds
one tied continued-fraction level, ``P(x) + softplus(beta)/(state + epsilon)``.
The matched contr... | {
"id": "2f20994f-f203-4cb0-b3dc-52993e1cbdd5",
"created_at": "2026-08-10 20:25:47.501642+00:00",
"db_md5": "60c097b303e1d8abc2a017917010a304",
"submitter": "Yash Kant",
"github_login": "yashkant",
"run_id": "087b0975-74e6-4b6a-b3ce-c7e4851edc19",
"tier": "easy",
"dataset_id": "e5",
"status": "succeed... | {
"score": {
"mean_loss": 2.150332898138269,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.009583333345750967
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.1564353100387503,
"example_count": 600... |
2f283bd7-52ae-4707-bef9-ad563162b9d1 | easy | lpbb | 2026-08-07 13:27:33.671974+00:00 | succeeded | 51889b25a17f5926d88db98bdd04967401aff2383d4a8d82db10ee581b93eed5 | 18,068 | null | """Exhaustive lever map: every knob the competition API exposes, spelled out.
Behaves EXACTLY like submissions/dev/submission.py (same model, same AdamW,
every optional lever left at its neutral default) — but every lever is written
out explicitly with its constraints, so this file doubles as the reference for
"what a... | {
"id": "2f283bd7-52ae-4707-bef9-ad563162b9d1",
"created_at": "2026-08-07 13:27:33.671974+00:00",
"db_md5": "e3070b71ac000068eca7258346672de7",
"submitter": "Lpbb",
"github_login": "lpbb",
"run_id": "3a49547f-7f25-4559-82ac-690cd2528c16",
"tier": "easy",
"dataset_id": "e3",
"status": "succeeded",
"s... | {
"score": {
"mean_loss": 13.862245888808712,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0068750000279396775
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 14.398883696007639,
"example_count": 8... |
2f2f3567-d55a-4f87-8ded-b24c3feb5d2d | easy | 0Chris5R | 2026-08-08 11:36:26.229940+00:00 | succeeded | e35d8bf215d43cf3fc7df1effe6a75d3c7186d40e00159ee03f89394cb914931 | 21,029 | null | """Shared-digit Neural GPU with a compute-matched operator curriculum.
The model aligns decimal digits by significance, forms a learned low-rank
second-order feature map of the current state, and evolves a two-dimensional
workspace with the original convolutional-GRU equations. Each arithmetic
transition applies the ... | {
"id": "2f2f3567-d55a-4f87-8ded-b24c3feb5d2d",
"created_at": "2026-08-08 11:36:26.229940+00:00",
"db_md5": "f3b18acd14a796d01b5c730e31e33563",
"submitter": "Chris ",
"github_login": "0Chris5R",
"run_id": "40a31d23-b353-4108-9bba-5f6ae3d02921",
"tier": "easy",
"dataset_id": "e5",
"status": "succeeded"... | {
"score": {
"mean_loss": 3.0791612143489426,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.03166666666666667
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.9903025252936666,
"example_count": 600... |
2f2f5916-daa9-437b-91c1-9d34eb858dc6 | easy | mnida | 2026-08-23 21:57:53.662610+00:00 | succeeded | 4de2601a507f8cf1756e8262690dd9fcb5fc8207904e3d2d546e528017ae71d1 | 7,406 | null | """Basic single-pass Transformer with PyTorch AdamW."""
from __future__ import annotations
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
TokenLossBatch,
assert_model_state,
)
D_MODEL =... | {
"id": "2f2f5916-daa9-437b-91c1-9d34eb858dc6",
"created_at": "2026-08-23 21:57:53.662610+00:00",
"db_md5": "f9ff2d1983300944ab5009bad0a155f2",
"submitter": "mnida",
"github_login": "mnida",
"run_id": "e7698adb-6326-4faf-9c14-63ba576cab46",
"tier": "easy",
"dataset_id": "e6",
"status": "succeeded",
... | {
"score": {
"mean_loss": 5.069290637969971,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.03851351607590914
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 4.632532119750977,
"example_count": 60,
... |
2f350599-1743-4aa1-a5d7-4eae3bc11d17 | easy | KaustubhKumar05 | 2026-08-26 03:10:06.232384+00:00 | succeeded | 78b50b2d9c1f1d8a40058f66faf508d698adec4df3febd10a9170f67b666e23f | 14,833 | null | """Neural-GPU-style conv GRU, v2s: v2 + speedrun pack.
- LR: warmup, hold max to 85% of wall-clock, steep late cliff (2026 schedule theory)
- identity-at-init: cell output projection zero-initialized
- batch reuse x6 (multi-pass theory)
Base (v2):
Changes over v1 (each backed by measured results in the literature):
... | {
"id": "2f350599-1743-4aa1-a5d7-4eae3bc11d17",
"created_at": "2026-08-26 03:10:06.232384+00:00",
"db_md5": "ea81573c182999541f4bee01aab78962",
"submitter": "koz",
"github_login": "KaustubhKumar05",
"run_id": "cf28171b-bf1b-4a0c-bea2-ad2fafa1c4ab",
"tier": "easy",
"dataset_id": "e6",
"status": "succee... | {
"score": {
"mean_loss": 2.883633852005005,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.5324324667453766
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 1.3764123916625977,
"example_count": 60,
... |
2f356fba-aece-4f84-8658-5ef1de4e112e | easy | erdavis0 | 2026-08-24 09:31:53.719190+00:00 | succeeded | aff7ab9c45ff330dc6a9a3134ae6fb56a6103591ee1fa658e9077c4e6664301d | 19,176 | null | """Canonical-reset learned macro digit transducer.
Public delimiter syntax is used only to route opaque token identities onto
separate right-aligned tapes. One T-blind learned macro maps a categorical X
tape and a digitwise N tape to a new categorical tape through freshly reset
interaction, carry, and modulus-conditi... | {
"id": "2f356fba-aece-4f84-8658-5ef1de4e112e",
"created_at": "2026-08-24 09:31:53.719190+00:00",
"db_md5": "c6fd794775990d735bc817407525b8c6",
"submitter": "Ethan",
"github_login": "erdavis0",
"run_id": "31995186-a8e0-40ae-94f6-8bbcc1742d4e",
"tier": "easy",
"dataset_id": "e9",
"status": "succeeded",... | {
"score": {
"mean_loss": 1.8583301305770874,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.013333333656191826
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 1.8629794120788574,
"example_count": 90... |
2f3572b6-f3da-4e07-898f-058af5223617 | easy | Bananafly | 2026-08-12 11:57:36.318577+00:00 | succeeded | 809fbc2ed83e8f6656f33764c186ca181b6e92f471835762cfdafa750937f01c | 16,610 | null | """T-aligned latent register automaton for the Easy E5 architecture screen."""
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,
asser... | {
"id": "2f3572b6-f3da-4e07-898f-058af5223617",
"created_at": "2026-08-12 11:57:36.318577+00:00",
"db_md5": "abc584851a33c79cb9140dab4268e751",
"submitter": "Andre Kreidemann",
"github_login": "Bananafly",
"run_id": "78ecc905-0375-49ae-a01f-e2872c788072",
"tier": "easy",
"dataset_id": "e5",
"status": ... | {
"score": {
"mean_loss": 4.395795188844204,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.007083333426465591
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 4.548173427581787,
"example_count": 600,... |
2f35b1b0-b989-468b-93a7-4bcfe5def171 | easy | GLauzza | 2026-08-22 17:06:59.122172+00:00 | succeeded | 34b2155bfb56a548203499ffec976e47351ebc80ab043e5e37132e932efdce5b | 4,687 | null | """Basic single-pass Transformer with PyTorch AdamW."""
from __future__ import annotations
import math
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
assert_model_state,
)
D_MODEL = 16
NUM_... | {
"id": "2f35b1b0-b989-468b-93a7-4bcfe5def171",
"created_at": "2026-08-22 17:06:59.122172+00:00",
"db_md5": "b25d69ee354be2e394e5c951a7877103",
"submitter": "Gabriel",
"github_login": "GLauzza",
"run_id": "b41bf5e7-f777-48ea-b70d-a05143cadeee",
"tier": "easy",
"dataset_id": "e1",
"status": "succeeded"... | {
"score": {
"mean_loss": 13.095979690551758,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0033333334140479565
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 14.087382316589355,
"example_count": 1... |
2f368268-daf4-4d3b-ab8a-be1228994caf | easy | AFeinfield | 2026-08-18 23:17:23.489037+00:00 | succeeded | f3ea9def11aed7616ceb8c63bade2f500ba750d00867faa661a9e92712f4d71e | 23,066 | null | """078: DEPTH CURRICULUM on top of the 077 active-row trajectory loop.
The T-indexed loop costs one block application per squaring, so a row with
T=16 is sixteen times as expensive as a row with T=1. On the medium tiers
(T drawn from {4, 8, 16}) the deep rows dominate the step time while the
one-step map they all sha... | {
"id": "2f368268-daf4-4d3b-ab8a-be1228994caf",
"created_at": "2026-08-18 23:17:23.489037+00:00",
"db_md5": "96b7c1cbc9c4ae9f6495656fc90afe27",
"submitter": "AFeinfield",
"github_login": "AFeinfield",
"run_id": "06b4f171-309c-4c82-ac9a-6d7a8a7acb04",
"tier": "easy",
"dataset_id": "e2",
"status": "succ... | {
"score": {
"mean_loss": 3.061005115509033,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.008125000167638063
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.529646158218384,
"example_count": 300,... |
2f589924-df3b-4c40-832d-a10cd696cf55 | easy | karanganesan | 2026-08-05 10:20:07.586462+00:00 | succeeded | 0f43aa1d3caa0d19c00c6be8f6f87c23391e53d75b08f4cc94ce89e93aebd09e | 22,080 | 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": "2f589924-df3b-4c40-832d-a10cd696cf55",
"created_at": "2026-08-05 10:20:07.586462+00:00",
"db_md5": "9579cf164b871ce8e59d031cdc46354f",
"submitter": "Karan Ganesan",
"github_login": "karanganesan",
"run_id": "36ef3c87-9228-44a4-a79f-3fa8d98e1fe0",
"tier": "easy",
"dataset_id": "e1",
"status": ... | {
"score": {
"mean_loss": 8.202332258224487,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.018333333544433117
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 7.2174153327941895,
"example_count": 100... |
2f60c293-3b41-4753-8c13-5cda00996f1a | easy | DDanlov | 2026-08-27 22:06:24.773193+00:00 | succeeded | 5b85ac8a83a4e5610c4d4caa93dac757900d9478cf131b0fd9c142d0092e8e33 | 39,442 | null | import math
import time
from dataclasses import dataclass
from typing import Optional, Tuple, Dict, Any, List, Union
import torch
import torch.nn as nn
import torch.nn.functional as F
from torch.optim import Optimizer
try:
from onelayerdeeper.api import Submission, OptimizerBundle, ModelSpec, OptimizerSpec
except ... | {
"id": "2f60c293-3b41-4753-8c13-5cda00996f1a",
"created_at": "2026-08-27 22:06:24.773193+00:00",
"db_md5": "c367c506519ab3a06e10c5aa3ad83bd5",
"submitter": "DDanlov",
"github_login": "DDanlov",
"run_id": "11695561-a455-4365-8d93-a52b09f6989f",
"tier": "easy",
"dataset_id": "e5",
"status": "succeeded"... | {
"score": {
"mean_loss": 2.7383499145507812,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0025000000605359674
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.740278482437134,
"example_count": 60... |
2f62ff43-7245-4407-ab96-105bc775b60f | easy | sapient-sapiens | 2026-08-20 17:37:35.525589+00:00 | succeeded | 413769f5ba6a90e8b895eef1b1958be778fffe9a9defa3ed7b69ffa304dab114 | 44,919 | null | """Persistent pair-indexed relational state around a categorical operator.
The recurrent answer state remains a bank of right-aligned categorical digit
slots. A separate continuous S-by-S grid retains every ordered digit-pair
cell across applications. Learned bilinear bindings and categorical modulus
context update ... | {
"id": "2f62ff43-7245-4407-ab96-105bc775b60f",
"created_at": "2026-08-20 17:37:35.525589+00:00",
"db_md5": "cc9a7e0598424673ad32472fa8a6ab97",
"submitter": "Ertondy",
"github_login": "sapient-sapiens",
"run_id": "b149477c-2e6c-40dd-95f9-d8d2cc776b8c",
"tier": "easy",
"dataset_id": "e5",
"status": "su... | {
"score": {
"mean_loss": 2.2570503905941677,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.009166666666666667
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.2555417661709636,
"example_count": 60... |
2f65a57f-c2ec-4c16-bdc2-6d05f23233a7 | easy | DDanlov | 2026-08-22 23:48:05.600427+00:00 | succeeded | 3f04351c3325914cd7789d27f4880b860dc25e924542c95e841fdc1e091ebf2f | 28,852 | 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": "2f65a57f-c2ec-4c16-bdc2-6d05f23233a7",
"created_at": "2026-08-22 23:48:05.600427+00:00",
"db_md5": "4c4b5dbeb280dfd4222e3c5e62821ad1",
"submitter": "DDanlov",
"github_login": "DDanlov",
"run_id": "19175ea6-c209-4947-8e8a-c563f4e4404a",
"tier": "easy",
"dataset_id": "e5",
"status": "succeeded"... | {
"score": {
"mean_loss": 259.4891778739184,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.005
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 253.69329149733744,
"example_count": 600,
"ex... |
2f672d2c-b179-4d77-94ea-954710cf80cb | easy | viridale | 2026-08-26 00:13:46.239746+00:00 | succeeded | fd0e864ac73f079db7a9b6c883add6e230a2e53122eec447f44e58a84eede8b4 | 18,830 | null | """Progressive residual pair Transformer.
hypothesis: v919 instantiates but its auxiliary exit remains in full input-sequence
coordinates while TokenLossBatch labels occupy target slots. Gather the auxiliary
exit through evaluator-supplied target_positions; prediction semantics are unchanged
axis: arch/loss
ta... | {
"id": "2f672d2c-b179-4d77-94ea-954710cf80cb",
"created_at": "2026-08-26 00:13:46.239746+00:00",
"db_md5": "de66b3296ab4514d41d5fb6cf8a242e9",
"submitter": "priormancer",
"github_login": "viridale",
"run_id": "994406b3-5652-44d3-9599-0c37ad5d6bee",
"tier": "easy",
"dataset_id": "e6",
"status": "succe... | {
"score": {
"mean_loss": 2.571893632411957,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.24594596028327942
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 1.915706753730774,
"example_count": 60,
... |
2f6a75ed-afb0-4787-932f-6a9c46b92e4b | easy | yashkant | 2026-08-15 14:31:49.805991+00:00 | succeeded | 2a920c54a3b75021f26183100b2c3c8babdcdc24886715ec4e1a11e6be85ad1c | 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": "2f6a75ed-afb0-4787-932f-6a9c46b92e4b",
"created_at": "2026-08-15 14:31:49.805991+00:00",
"db_md5": "b3bf544ca21f3ce76ceb8742d748c33a",
"submitter": "Yash Kant",
"github_login": "yashkant",
"run_id": "b7f294f0-0602-4e21-ad74-fa304c54c496",
"tier": "easy",
"dataset_id": "e4",
"status": "succeed... | {
"score": {
"mean_loss": 2.06306219482017,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.009375000024835269
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.06577878017359,
"example_count": 1200,
... |
2f7940e0-08a7-4567-8049-fea84b06ca71 | easy | Cree0618 | 2026-08-30 13:53:38.161254+00:00 | failed | c2fb2338d1a8b67380a83af382ac7c406ba969e34e6572cb0a4f14a8143b583f | 8,592 | null | """E047: bit-serial NeuralHorner challenger for modular squaring."""
from __future__ import annotations
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import ModelSpec, OptimizerBundle, OptimizerSpec, Submission, TokenLossBatch, assert_model_state
PAD, X, T, ANS, EOS, OFF = 0... | {
"id": "2f7940e0-08a7-4567-8049-fea84b06ca71",
"created_at": "2026-08-30 13:53:38.161254+00:00",
"db_md5": "efbbdcbc37dc232e373b793a0d77bf19",
"submitter": "Cree0618",
"github_login": "Cree0618",
"run_id": "3dcaab6c-55fc-4f76-89c8-0b1703e7f29c",
"tier": "easy",
"dataset_id": "e5",
"status": "failed",... | null
|
2f7fffd6-6e51-4400-bf9d-04a0b47d2d48 | easy | yunjiangster | 2026-08-27 18:44:13.086383+00:00 | succeeded | 1834dca679e1bd0aaec7cd276e7647dcabdeeb05dff7f0afad510809663a400b | 10,897 | null | """Significance-tape Deep Thinking probe for One Layer Deeper.
This is a generic learned recurrent solver. The only deterministic structure is
prompt parsing and alignment of decimal places. All arithmetic behavior is
learned from random initialization.
"""
from __future__ import annotations
import math
import time... | {
"id": "2f7fffd6-6e51-4400-bf9d-04a0b47d2d48",
"created_at": "2026-08-27 18:44:13.086383+00:00",
"db_md5": "9a26cbc512bf1b915bdf93ec33a5e62a",
"submitter": "Yunjiang Jiang",
"github_login": "yunjiangster",
"run_id": "d0867334-4dd2-42f4-89b4-f6b7d4131299",
"tier": "easy",
"dataset_id": "e1",
"status":... | {
"score": {
"mean_loss": 2.5061380863189697,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.036666667088866234
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.6221277713775635,
"example_count": 10... |
2f805474-562c-4161-89bb-7793414cec60 | easy | himalalps | 2026-08-07 16:04:49.488661+00:00 | succeeded | 174130f2b6ed200569ec004ce3c5d8f67bd460875d8cd6d47f7b7cd744f6d8ce | 69,778 | 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": "2f805474-562c-4161-89bb-7793414cec60",
"created_at": "2026-08-07 16:04:49.488661+00:00",
"db_md5": "c6e19bb64b5fd6bea2c5aa2239fd42e0",
"submitter": "Haoyu Tang",
"github_login": "himalalps",
"run_id": "e2de862e-1853-42cf-b90c-74989bcfab81",
"tier": "easy",
"dataset_id": "e5",
"status": "succe... | {
"score": {
"mean_loss": 3.2753852632668643,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.013750000198682149
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.243236780166626,
"example_count": 600... |
2f86cd9f-5dad-4a25-a1b8-9ab1438fecd5 | easy | blake-camp-surge | 2026-08-05 18:33:33.634524+00:00 | succeeded | e774a170475ba2cf49c7276d8ba197b5632425533e88abbd64981571edeeebf1 | 19,135 | null | """Width-16 linear-memory DAN with differentiable scratch nodes."""
from __future__ import annotations
import math
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
TokenLossBatch,
assert_m... | {
"id": "2f86cd9f-5dad-4a25-a1b8-9ab1438fecd5",
"created_at": "2026-08-05 18:33:33.634524+00:00",
"db_md5": "5e8ca7f638c66ad7f9bc6d8b81e58c40",
"submitter": "Blake Camp",
"github_login": "blake-camp-surge",
"run_id": "5248c74f-c0c4-44ef-a914-1bc18dc94ef6",
"tier": "easy",
"dataset_id": "e3",
"status":... | {
"score": {
"mean_loss": 2.1302404804389607,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.008125000009313226
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.127675242317691,
"example_count": 800... |
2f899759-4c24-481b-87b6-f9ab44fe2616 | easy | karanganesan | 2026-08-07 09:00:52.715370+00:00 | succeeded | 78d822b7d1f9bb5b46f977e12ceb542f196ddd50f2f7a90f51b4f9012f406971 | 25,547 | null | """Parametric looped-transformer family (P1).
One weight-tied transformer block applied k times in latent space. Config
flags cover four P1 families with one file:
looped recall=0 gated=0 tfilm=0 plain weight-tied loop
looped-recall recall=1 re-inject the input embedding each
... | {
"id": "2f899759-4c24-481b-87b6-f9ab44fe2616",
"created_at": "2026-08-07 09:00:52.715370+00:00",
"db_md5": "c6174b599773d987b1c6e74c1c2f0956",
"submitter": "Karan Ganesan",
"github_login": "karanganesan",
"run_id": "2ececb6a-531f-4537-9aa3-b43b795e5a40",
"tier": "easy",
"dataset_id": "e5",
"status": ... | {
"score": {
"mean_loss": 7.11957573890686,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.007916666800156236
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 7.196868896484375,
"example_count": 600,
... |
2f8ba84e-353d-4d11-8627-3da563673ed1 | easy | Maxwells-Demons | 2026-08-13 08:56:05.208842+00:00 | succeeded | 6abbebeebcc3863c286eb320bc9e9af440ef4451af836d64d9c3174be89853de | 109,857 | null | """One shared digit operator, its two free neighbours of x, and two learned maps.
The backbone is unchanged from v47/v96: the state IS the digit one-hot, one
application of the operator is
(state digits, N digits) -> next state digit logits
and the straight-through code of those logits is the whole state handed to... | {
"id": "2f8ba84e-353d-4d11-8627-3da563673ed1",
"created_at": "2026-08-13 08:56:05.208842+00:00",
"db_md5": "dab8b95432da3f12736d74330e25643d",
"submitter": "Maxwells_Ayakashi",
"github_login": "Maxwells-Demons",
"run_id": "8a7fcff9-f372-48f0-97d7-1b59ce2fc6af",
"tier": "easy",
"dataset_id": "e5",
"st... | {
"score": {
"mean_loss": 12.77778382424354,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.49333333353201547
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 7.639461519472268,
"example_count": 600,
... |
2f90bbf7-3e77-41d0-9344-084094c54383 | easy | pepinu | 2026-08-19 06:38:19.507124+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": "2f90bbf7-3e77-41d0-9344-084094c54383",
"created_at": "2026-08-19 06:38:19.507124+00:00",
"db_md5": "d82287d6697a27328f0e1b35067c619a",
"submitter": "pepinu",
"github_login": "pepinu",
"run_id": "14859a90-a9bd-4112-90f0-079997e08110",
"tier": "easy",
"dataset_id": "e3",
"status": "succeeded",
... | {
"score": {
"mean_loss": 24.61076997000541,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.5962499982118606
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 29.805143719818794,
"example_count": 800,
... |
2f91f9c0-bcf0-4675-a587-8a43f39a5d16 | easy | jordanrubin | 2026-08-11 01:23:53.193689+00:00 | succeeded | 67982e9f505a2cf74a8823191467477f81e0c11b7660f6c116f0fafa565fb3eb | 21,669 | 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": "2f91f9c0-bcf0-4675-a587-8a43f39a5d16",
"created_at": "2026-08-11 01:23:53.193689+00:00",
"db_md5": "5f41f74e5ab605af5d290f14916e4deb",
"submitter": "Jordan Rubin",
"github_login": "jordanrubin",
"run_id": "15947fce-424a-4d56-984a-a0d5ceb1fe46",
"tier": "easy",
"dataset_id": "e5",
"status": "s... | {
"score": {
"mean_loss": 2.387097360316355,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.00874999980442226
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.156008034142678,
"example_count": 600,
... |
2f922dac-909b-4bc9-9a0e-5c5d7c299966 | easy | Yalyenea | 2026-08-11 20:24:26.238587+00:00 | succeeded | 670b7e15bdf6bb92040e4f6606aac6c31a0f979303b0c8dd9f29a60ee6a5e31e | 15,057 | null | """R4 full-bandwidth candidate with fast causal field-local decimal registers."""
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_s... | {
"id": "2f922dac-909b-4bc9-9a0e-5c5d7c299966",
"created_at": "2026-08-11 20:24:26.238587+00:00",
"db_md5": "8a9835a98d8b851fe193b9c8a611903f",
"submitter": "yfff",
"github_login": "Yalyenea",
"run_id": "907396c1-6769-4686-9ac1-49224235cc9e",
"tier": "easy",
"dataset_id": "e3",
"status": "succeeded",
... | {
"score": {
"mean_loss": 8.54489517211914,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0118749993853271
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 8.458154678344727,
"example_count": 800,
... |
2fa890c6-32ec-4294-a68e-0b4e41e0917b | easy | viridale | 2026-08-28 11:42:29.690572+00:00 | succeeded | 02734c0cfe811b709c8c914eda79f9d432efdb188b4946d5332bada2dbc75d7a | 32,233 | null | """Dynamic-horizon shallow-evidence absolute learned coefficient slab.
hypothesis: v975's selected learned program executes all 64 transitions for every
train/eval batch, even though tensor masking makes every transition above max(T) an
exact no-op. Its complete candidate path already elides the same masked su... | {
"id": "2fa890c6-32ec-4294-a68e-0b4e41e0917b",
"created_at": "2026-08-28 11:42:29.690572+00:00",
"db_md5": "5979b4e4e41f59e41dbdd5d87806a165",
"submitter": "priormancer",
"github_login": "viridale",
"run_id": "03c2aec0-e79f-4b8d-9319-718f94499f6f",
"tier": "easy",
"dataset_id": "e5",
"status": "succe... | {
"score": {
"mean_loss": 27.944037034040278,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.5041666667039195
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 27.847533521096267,
"example_count": 600,... |
2faf8a06-ba6e-4a50-9790-0c31b0e8720f | easy | Yalyenea | 2026-08-11 21:51:14.539509+00:00 | succeeded | a70abb21410324c0cf167932a74c778826fd48c0a553813f711b0b5a55667788 | 12,845 | null | """R4 full-bandwidth candidate with a learned scratch-state readout."""
from __future__ import annotations
import math
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
assert_model_state,
)
D... | {
"id": "2faf8a06-ba6e-4a50-9790-0c31b0e8720f",
"created_at": "2026-08-11 21:51:14.539509+00:00",
"db_md5": "d2ce47dfe00b550a29bcc443b4a4a38b",
"submitter": "yfff",
"github_login": "Yalyenea",
"run_id": "80aa4e2a-8d99-4d8c-ad6f-bbadcc4bb8fb",
"tier": "easy",
"dataset_id": "e5",
"status": "succeeded",
... | {
"score": {
"mean_loss": 5.9997970153082605,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.011250000540167093
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 5.814354419708252,
"example_count": 600... |
2fb1c10f-bbf7-41fd-99ae-c42331124887 | easy | dlyr3 | 2026-08-23 19:44:01.131678+00:00 | failed | f94b635f6dd2b70c17e8a3d599ceaea226881d1a6201c012187cb986cfacb760 | 7,386 | null | """The M-layer of arXiv:2008.03936 applied to the repeated-squaring task.
M = B + T(U x) is built from the pooled prompt, exp(M) is taken with
torch.matrix_exp, and V + S exp(M) is read back out. The paper's activity regularisation is deliberately NOT applied: measured on e1,
any nonzero lambda drives ||exp(M)||_F... | {
"id": "2fb1c10f-bbf7-41fd-99ae-c42331124887",
"created_at": "2026-08-23 19:44:01.131678+00:00",
"db_md5": "3f569f9edd87fb3629c5fd5274b314cf",
"submitter": "dlyr3",
"github_login": "dlyr3",
"run_id": "64cce937-0a38-4932-85d3-084c4d578980",
"tier": "easy",
"dataset_id": "e1",
"status": "failed",
"st... | null
|
2fbdcbdd-a2c2-405d-b19c-883ed9baaba1 | easy | chad-atexpedient | 2026-08-10 02:07:05.385099+00:00 | succeeded | a27fc5f26227824f0af93a883ef603efe5a0e87576d3b5c38b100c55d9d77c64 | 10,934 | null | """BL8B: width-32 trajectory readout."""
from __future__ import annotations
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
assert_model_state,
)
LANES = 6
IMMUTABLE_LANES ... | {
"id": "2fbdcbdd-a2c2-405d-b19c-883ed9baaba1",
"created_at": "2026-08-10 02:07:05.385099+00:00",
"db_md5": "0987645d491e4a48ff663a109e36852e",
"submitter": "chad-atexpedient",
"github_login": "chad-atexpedient",
"run_id": "0cee7330-cb3d-478c-892c-1afbdd0387b5",
"tier": "easy",
"dataset_id": "e5",
"st... | {
"score": {
"mean_loss": 5.893434685840465,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0033333333333333335
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 5.884266878991919,
"example_count": 600... |
2fc16b3e-b3a2-4c04-b879-514052123ec8 | easy | DDanlov | 2026-08-10 09:34:05.580865+00:00 | failed | b0eb0d0456f4944fe2209594c0192cb39f6a86db7a33ae7a371012220dbdf5af | 16,175 | 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 RMSNorm(nn.Module):
def __init__(self, dim: int, eps: float = 1e-6):
super... | {
"id": "2fc16b3e-b3a2-4c04-b879-514052123ec8",
"created_at": "2026-08-10 09:34:05.580865+00:00",
"db_md5": "606904cbc90f06bc6b13ae6eef33f344",
"submitter": "DDanlov",
"github_login": "DDanlov",
"run_id": "9c276559-d653-49ed-9e5b-0b0bc2c51193",
"tier": "easy",
"dataset_id": "e1",
"status": "failed",
... | null
|
2fc563a4-be34-4411-b933-bfaaefb8ebf4 | easy | alirezashirvani-jr | 2026-08-31 23:06:20.587884+00:00 | succeeded | 833072d99b482a874a60ee8cfe1cca4b839a6255660346ef90e5069b04df0f06 | 25,284 | null | from __future__ import annotations
import math
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
TokenLossBatch,
assert_model_state,
)
HP = {
"width": 0,
"maximum_outer_steps": 64,... | {
"id": "2fc563a4-be34-4411-b933-bfaaefb8ebf4",
"created_at": "2026-08-31 23:06:20.587884+00:00",
"db_md5": "419d665ac4de16d355ed6fa3232bf1da",
"submitter": "alirezashirvani-jr",
"github_login": "alirezashirvani-jr",
"run_id": "e0a1a376-a235-4807-b94f-afc9014227d5",
"tier": "easy",
"dataset_id": "e5",
... | {
"score": {
"mean_loss": 2.6844997506499753,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.002500000024835269
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.6833491111550094,
"example_count": 60... |
2fc7430d-5e7e-4374-b40b-a29412ca7db6 | easy | AFeinfield | 2026-08-27 22:35:38.276339+00:00 | succeeded | 6919b57aaf377835920efffa83a797cf7452c218b452093e714f48c7a4ed5f00 | 13,402 | null | """158: IDENTITY-BIASED OPERATORS — escape the marginals basin.
155/156/157 all converged to the same shallow attractor (loss ~1.87,
T=1 rung ~0, deep-T rungs inflated by orbit fixed points). Fixes, all
generic (random weights, no task values):
- CarrySplit: net emits only carry (small-init ~0); digit = y - 10*carry
... | {
"id": "2fc7430d-5e7e-4374-b40b-a29412ca7db6",
"created_at": "2026-08-27 22:35:38.276339+00:00",
"db_md5": "2542f32a208390658e377ca42dc02164",
"submitter": "AFeinfield",
"github_login": "AFeinfield",
"run_id": "0237e4c3-5af1-4978-bb22-75e3dcad2add",
"tier": "easy",
"dataset_id": "e1",
"status": "succ... | {
"score": {
"mean_loss": 390.1333614587784,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.06833333522081375
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 778.1959228515625,
"example_count": 100,
... |
2fce10ae-df5a-4cd6-8e9a-7e10aaf2d2e1 | easy | sapient-sapiens | 2026-09-01 02:50:22.204961+00:00 | succeeded | bb67ea8545822e2265657b0289b989bb1f329d79ad5bd6168fed91d9591d7304 | 24,390 | null | """E5: a tested one-shot D6 square tied only across task time T.
The physical Transformer block is one complete learned square operator. It is
called once per square boundary and receives no raw N token: categorical N
only edits the shared weights and character table. Eight boundary calls are
fully differentiable; a... | {
"id": "2fce10ae-df5a-4cd6-8e9a-7e10aaf2d2e1",
"created_at": "2026-09-01 02:50:22.204961+00:00",
"db_md5": "c557e652ee214d6ae97e0ca783cbb7c4",
"submitter": "Ertondy",
"github_login": "sapient-sapiens",
"run_id": "2bc8e7fa-8248-477b-91ed-5a56b7d1548a",
"tier": "easy",
"dataset_id": "e5",
"status": "su... | {
"score": {
"mean_loss": 2.1460313337987067,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.004583333333333333
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.1462606591226807,
"example_count": 60... |
2fd1d49a-f36a-4bca-9ef4-0d28cb05c363 | easy | blake-camp-surge | 2026-08-04 15:54:59.244652+00:00 | succeeded | 2226542a37d738b72dd9354867bd8f0bd95f1c47edf45a78ee087681581f001f | 14,969 | null | """Vector DANs communicating through recurrent linear-attention memory."""
from __future__ import annotations
import math
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
TokenLossBatch,
a... | {
"id": "2fd1d49a-f36a-4bca-9ef4-0d28cb05c363",
"created_at": "2026-08-04 15:54:59.244652+00:00",
"db_md5": "9bf932f01ba9c425f5f9286668849be5",
"submitter": "Blake Camp",
"github_login": "blake-camp-surge",
"run_id": "a79a39a7-8d3a-4c93-9017-aa6991730308",
"tier": "easy",
"dataset_id": "e3",
"status":... | {
"score": {
"mean_loss": 2.1541987695919147,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.01625000015832484
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.153572580620778,
"example_count": 800,... |
2fd70220-87a1-48fd-89b1-3da0cbb72e67 | easy | viridale | 2026-08-02 09:41:52.230614+00:00 | succeeded | aa9d65ae04736b19b24fa42a1c692e90d10fc8aed56b20bd2b1bed4a60fe7f24 | 31,692 | null | """Hybrid exact-program route plus exact-T=1 active transition learning.
hypothesis: v087 proves a catalogue-open table can learn a recurrence whose
best one-edit grammar atom is only 7.43% correct, while preserving official
e1/e3/m1 64/64. Its m1 custom loss nevertheless remains near 3 nats because
a targ... | {
"id": "2fd70220-87a1-48fd-89b1-3da0cbb72e67",
"created_at": "2026-08-02 09:41:52.230614+00:00",
"db_md5": "c38511626fc054e8a5be6f4f5c530897",
"submitter": "priormancer",
"github_login": "viridale",
"run_id": "660bdb93-7c31-4a4a-b0a5-1ad9f11aaf62",
"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,
"... |
2fd79e0b-47f3-4f73-9155-07939ab3c1bc | easy | sapient-sapiens | 2026-08-17 09:26:40.228721+00:00 | succeeded | 2ea1f2434f96ff4852b5a0a0f551298535e47ea33c39caef76542ebd44ab9d80 | 17,289 | 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": "2fd79e0b-47f3-4f73-9155-07939ab3c1bc",
"created_at": "2026-08-17 09:26:40.228721+00:00",
"db_md5": "bb0020f7151a9b150e0e4a0c0a9b16a0",
"submitter": "Ertondy",
"github_login": "sapient-sapiens",
"run_id": "f46984d7-f1ec-4a11-a8fb-bcd1559fa739",
"tier": "easy",
"dataset_id": "e3",
"status": "su... | {
"score": {
"mean_loss": 3.0146462584131695,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.001875
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 3.0146211279907997,
"example_count": 800,
... |
2fd9dbed-3a29-4135-ab2f-dc60442e291e | easy | rougee | 2026-08-23 16:29:27.589588+00:00 | failed | ed1e8386966eb696f6b9e8ff96fa71b4289ee4e0ee2a0b92db6df15b8c757a54 | 13,161 | null | """Official Submittable Champion for One Layer Deeper (Matrix-Exponential SO(d) + Dyadic Harmonic Core).
Hardware & Algebraic Co-Design:
1. Dyadic Harmonic Phasor Prelude: Power-of-2 spectrum omega_k = 2^k * pi on (S^1)^16 (pointwise, zero dim=1 reduction)
2. Matrix-Exponential Lie Group SO(d): W = exp(S) where S = A ... | {
"id": "2fd9dbed-3a29-4135-ab2f-dc60442e291e",
"created_at": "2026-08-23 16:29:27.589588+00:00",
"db_md5": "48e86b270916338e97de161fd3330684",
"submitter": "Rougee",
"github_login": "rougee",
"run_id": "0d945fc1-a7ec-4a24-a9b0-7d16bc75a6b8",
"tier": "easy",
"dataset_id": "e1",
"status": "failed",
"... | null
|
2fe53acc-119c-40c5-bb01-b7e55337c097 | easy | oupadhyay | 2026-08-12 06:45:52.035708+00:00 | succeeded | c405fee90dd2bb47e5513b3f740d107b850b9c8fd3799f98a6dc4b8b32b79de2 | 8,404 | null | """Standalone RMS/axial crossing C12 for the Easy E5 tier."""
from __future__ import annotations
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import ModelSpec, OptimizerBundle, OptimizerSpec, Submission, assert_model_state
PAD, N_MARK, X_MARK, T_MARK, DIGIT_BASE =... | {
"id": "2fe53acc-119c-40c5-bb01-b7e55337c097",
"created_at": "2026-08-12 06:45:52.035708+00:00",
"db_md5": "16d8c92388a098eac39784ced0c5fa44",
"submitter": "Ojasw Upadhyay",
"github_login": "oupadhyay",
"run_id": "604620d7-a97d-4bb4-be5a-5403a11aded9",
"tier": "easy",
"dataset_id": "e5",
"status": "s... | {
"score": {
"mean_loss": 2.3158367962668214,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.002916666666666667
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.2202758126194704,
"example_count": 60... |
2fe7a5f8-e7a9-45e0-ac5c-9a3c935f9fe7 | easy | yashkant | 2026-08-16 00:31:31.112101+00:00 | succeeded | 83de680f8c03e86b303fd643c5d3e1086a6328e1de964bb3160a0c55cd37aae2 | 42,377 | null | """Round1801 finite-ring spectral composition inside W12 Dykstra.
The complete Round1765 W12/1T carrier stays active. Each parsed-T cycle maps
the live answer state into a low-capacity real character bank, rotates answer
and immutable X modes by categorical-N phases, and composes answer and latent
frequencies with a ... | {
"id": "2fe7a5f8-e7a9-45e0-ac5c-9a3c935f9fe7",
"created_at": "2026-08-16 00:31:31.112101+00:00",
"db_md5": "13c92c04e860bfcfefb4b6f13bf18758",
"submitter": "Yash Kant",
"github_login": "yashkant",
"run_id": "56b2e7d1-d8fa-4a9a-8e7f-68aa1de6a69e",
"tier": "easy",
"dataset_id": "e3",
"status": "succeed... | {
"score": {
"mean_loss": 2.125317081301863,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.017500000158324838
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.124355067111486,
"example_count": 800,... |
2ff17376-cb58-447e-92ef-c805976c663f | easy | blake-camp-surge | 2026-08-05 17:27:57.403614+00:00 | succeeded | 1e5ec9698d1c034b49bc54e59ad48a413eb4aa5827507ffac3132985f729d9a7 | 9,778 | null | """Differentiable soft scratchpad with adaptive recurrent pre-deliberation."""
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_mod... | {
"id": "2ff17376-cb58-447e-92ef-c805976c663f",
"created_at": "2026-08-05 17:27:57.403614+00:00",
"db_md5": "618acec3ea49e3ad84843a79318a6326",
"submitter": "Blake Camp",
"github_login": "blake-camp-surge",
"run_id": "7c7282fb-e81a-44ad-9698-845c47346bdc",
"tier": "easy",
"dataset_id": "e3",
"status":... | {
"score": {
"mean_loss": 2.1729418706161763,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.013125000000000001
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.188774736578544,
"example_count": 800... |
2ff193d4-95e6-4a33-9fcc-3b0f3620cf89 | easy | yashkant | 2026-08-14 17:28:22.605711+00:00 | succeeded | ad864e90e0e6bda466f52143caf7a261cc31049ae2c920708a8434536212d9c1 | 36,732 | null | """Round1800 N-fibered quotient-product recursion inside W12 Dykstra.
The complete Round1765 W12/1T carrier remains active. Before every recurrent
projection, the live lifted state is treated as coefficients, squared by one
shared bilinear convolution, and reduced through a learned monic relation
derived only from ca... | {
"id": "2ff193d4-95e6-4a33-9fcc-3b0f3620cf89",
"created_at": "2026-08-14 17:28:22.605711+00:00",
"db_md5": "8f5d49b2bfd9bfc92b73cb12b3ca3cb9",
"submitter": "Yash Kant",
"github_login": "yashkant",
"run_id": "36649d9f-8e8c-4a65-9af8-bbeec5a7e482",
"tier": "easy",
"dataset_id": "e3",
"status": "succeed... | {
"score": {
"mean_loss": 2.1222228566238446,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.01562500013038516
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.1222232447284384,
"example_count": 800... |
2ff1a31b-7384-4107-8772-f890f66a8ca8 | easy | jordanrubin | 2026-08-02 18:22:23.428595+00:00 | succeeded | fc83380fd71a375ea72ea5c0f6a095229d7241bed721084eff1844b20f9e70c0 | 13,994 | null | """Bidirectional transformer, D=256 x 3 blocks, grokking-tuned AdamW.
Design notes:
- Inputs are padded to config.max_seq_len inside forward and logits sliced
back, so every training batch presents one static shape (compile/cudagraph
friendly, and uniform kernel shapes even in eager).
- Attention mask gets an iden... | {
"id": "2ff1a31b-7384-4107-8772-f890f66a8ca8",
"created_at": "2026-08-02 18:22:23.428595+00:00",
"db_md5": "8424ff3c2b53438cdfeb088f9343f339",
"submitter": "Jordan Rubin",
"github_login": "jordanrubin",
"run_id": "415f2a0a-427b-4ea7-b38b-b2d37403edb3",
"tier": "easy",
"dataset_id": "e1",
"status": "s... | {
"score": {
"mean_loss": 5.369694111521751,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.2083333246409893
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 7.858675106247859,
"example_count": 100,
... |
300430e2-66c5-464c-abbc-66d30453a981 | easy | pculaf | 2026-08-15 22:38:51.746189+00:00 | succeeded | 03cd0a29d1507b97fc8a4a9b5aa62a579909a7301af8bf68cb15b41af8b11368 | 10,442 | null | """Muon-optimized recurrent Transformer for the E6 activation/depth sweep."""
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,
Submissio... | {
"id": "300430e2-66c5-464c-abbc-66d30453a981",
"created_at": "2026-08-15 22:38:51.746189+00:00",
"db_md5": "f1d27f897dae84266bb97ff79c5beeb0",
"submitter": "Pavle Culafic",
"github_login": "pculaf",
"run_id": "baf517f5-94c7-4bc7-9d6f-e11a88d084c9",
"tier": "easy",
"dataset_id": "e6",
"status": "succe... | {
"score": {
"mean_loss": 1.906350553035736,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.021846847608685493
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 1.883166790008545,
"example_count": 60,
... |
3010e0ad-2912-4e73-b03a-0117ef13070f | easy | DESU-CLUB | 2026-08-24 11:08:48.877898+00:00 | succeeded | be14714a8983c6a541e1aa30b0e9b1f221493a543fcda07f8ab364cc8796158f | 10,168 | null | """Huginn-style looped Transformer whose iteration count is read from the T token.
prelude -> [core applied T times, input re-injected each step] -> coda
Forked from experiments/_baselines/adamw_transformer.py; the Block, RMSNorm and
optimizer are unchanged so that the only varied axis is the depth mechanism.
"""
fr... | {
"id": "3010e0ad-2912-4e73-b03a-0117ef13070f",
"created_at": "2026-08-24 11:08:48.877898+00:00",
"db_md5": "475e03d6c608505997a47f7af0689520",
"submitter": "Low Keng Hoong, Warren",
"github_login": "DESU-CLUB",
"run_id": "26b96c9b-e612-4684-9b36-4b195085d6a6",
"tier": "easy",
"dataset_id": "e6",
"sta... | {
"score": {
"mean_loss": 7.439334470931797,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.03851351574868769
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 7.025671005249023,
"example_count": 60,
... |
3014002a-13c1-4257-bf49-3f1271f906e7 | easy | shirvani-jr | 2026-09-01 00:51:50.458176+00:00 | succeeded | b317b19da85760bde75e4d27c457d49e66ac8dcca62bdf058148eb0af50f5c6b | 24,687 | 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
ANSWER_TOKEN = 5
END_TOKEN = 6
DIGIT_OFFSET = 7
FACTOR_COUNT = 12
OUTER... | {
"id": "3014002a-13c1-4257-bf49-3f1271f906e7",
"created_at": "2026-09-01 00:51:50.458176+00:00",
"db_md5": "9af79d78d1f2438fc32ea2e1db49817a",
"submitter": "Ali",
"github_login": "shirvani-jr",
"run_id": "9da73a7e-5220-4365-aaff-7f2b866fa816",
"tier": "easy",
"dataset_id": "e4",
"status": "succeeded"... | {
"score": {
"mean_loss": 3.1273713956794094,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0016666666666666668
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 3.135120928851775,
"example_count": 12... |
301a9167-d933-481c-a7d3-27b2a9450e13 | easy | yashkant | 2026-08-13 00:34:15.297137+00:00 | succeeded | 7c74d210aa672064f4f306c1fd5b05638c34d41cbd7b4018e446da28f4ea6f7d | 40,954 | null | """Round1786 evolutionary crossover of two controlled survivors.
The Round1780 W8/D2 carrier keeps its winning contrastive, exactly
antisymmetric learned set codes. The candidate additionally removes each
initial and recurrent Dykstra correction residual's coordinate mean before
the unchanged tanh bound, combining Ro... | {
"id": "301a9167-d933-481c-a7d3-27b2a9450e13",
"created_at": "2026-08-13 00:34:15.297137+00:00",
"db_md5": "6b2852edb9c63d59879036ff93b71fcd",
"submitter": "Yash Kant",
"github_login": "yashkant",
"run_id": "d7fe5f51-db66-424e-8a23-287e9a6c9f59",
"tier": "easy",
"dataset_id": "e4",
"status": "succeed... | {
"score": {
"mean_loss": 2.0711097967018484,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.016041666691501935
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.074151036985252,
"example_count": 120... |
301fd24a-e9d5-4bd5-9eeb-1005b78bfe78 | easy | chad-atexpedient | 2026-08-14 14:50:39.495939+00:00 | succeeded | 3832270fe8996e33064a02b0fb7eb218d49205a8b91d14597055edfbb34c7cd8 | 17,315 | null | """X18: E04 with compositional consistency loss.
Creates state-level supervision by matching T=2 intermediate states with T=1
final states when the T=2 output equals the T=1 input (string comparison only).
No arithmetic is computed - only evaluator-supplied values are compared.
"""
from __future__ import annotations
... | {
"id": "301fd24a-e9d5-4bd5-9eeb-1005b78bfe78",
"created_at": "2026-08-14 14:50:39.495939+00:00",
"db_md5": "a878fe29eab3ed7abd3eab2cb4685fe4",
"submitter": "chad-atexpedient",
"github_login": "chad-atexpedient",
"run_id": "fa347b56-64f9-4271-9634-97e78bd6cfe4",
"tier": "easy",
"dataset_id": "e5",
"st... | {
"score": {
"mean_loss": 2.1859304904937744,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.00541666685603559
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.197192668914795,
"example_count": 600,... |
30205d4d-d63f-42c8-bd5d-7bd70e58d851 | easy | khushidahi | 2026-08-16 12:21:25.430126+00:00 | succeeded | 0270b725091378013f61231bf4ad366455076c33dca6f9219a49056d4d2f419c | 21,753 | null | """R99: early-stopped continuous latent functional-root model.
This candidate is deliberately diagnostic. It trains on evaluator-owned final
labels at several depths so that one shared transition must compose. The model
never converts N or X to native numbers and never implements multiplication or
modular reduction.... | {
"id": "30205d4d-d63f-42c8-bd5d-7bd70e58d851",
"created_at": "2026-08-16 12:21:25.430126+00:00",
"db_md5": "adebbc4ad18612d2c7e6ecc36a03b290",
"submitter": "khushidahi",
"github_login": "khushidahi",
"run_id": "edde631c-df43-47f9-b702-2505518ad8ac",
"tier": "easy",
"dataset_id": "e3",
"status": "succ... | {
"score": {
"mean_loss": 2.7746058661297424,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.006250000013969839
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.73023845340005,
"example_count": 800,... |
302d0c62-a293-4a1b-9d2e-62cdf5b8c560 | easy | DDanlov | 2026-08-28 21:49:17.370772+00:00 | succeeded | 97184c5c9a6abd3fc9ea2b7f6bf9249f1bcc1bcbec7e7e1df0ccc2987a4224d9 | 14,934 | null | """
Trial 18: Model B: D=384, H=6, 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": "302d0c62-a293-4a1b-9d2e-62cdf5b8c560",
"created_at": "2026-08-28 21:49:17.370772+00:00",
"db_md5": "7cb388fe561cd421d9fa2bcb2f35fd65",
"submitter": "DDanlov",
"github_login": "DDanlov",
"run_id": "484ae935-6e8e-4484-b513-ecd1fb4ed0f5",
"tier": "easy",
"dataset_id": "e5",
"status": "succeeded"... | {
"score": {
"mean_loss": 2.83321353302519,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.8332135109356167,
"example_count": 600,
"exact... |
302e7320-6da6-429c-a51c-4526936eeef6 | easy | DDanlov | 2026-08-08 09:47:17.432832+00:00 | failed | e31c1a90702d978e5d15f025a430cf3d223b4e8e05aefb87269eb95a19482be9 | 13,315 | 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": "302e7320-6da6-429c-a51c-4526936eeef6",
"created_at": "2026-08-08 09:47:17.432832+00:00",
"db_md5": "c44db1d74acc64f8dde4c53b1ab7d505",
"submitter": "DDanlov",
"github_login": "DDanlov",
"run_id": "cdc9ab43-b98d-4214-ba5c-d42bba91feac",
"tier": "easy",
"dataset_id": "e1",
"status": "failed",
... | null
|
303b717a-f821-4329-a56b-c479fb763d39 | easy | Maxwells-Demons | 2026-08-14 05:24:36.307766+00:00 | succeeded | 39d19782a48458e9b92675994b85fad78096b9058e931d958b4c01aad3f0a243 | 120,067 | null | """One shared digit operator, its two free neighbours of x, and two learned maps.
The backbone is unchanged from v47/v96: the state IS the digit one-hot, one
application of the operator is
(state digits, N digits) -> next state digit logits
and the straight-through code of those logits is the whole state handed to... | {
"id": "303b717a-f821-4329-a56b-c479fb763d39",
"created_at": "2026-08-14 05:24:36.307766+00:00",
"db_md5": "5f70eb0891324e3a78ade8f41ff32bd6",
"submitter": "Maxwells_Ayakashi",
"github_login": "Maxwells-Demons",
"run_id": "f3222cb6-d6da-4727-8d97-5ed1bfa426ed",
"tier": "easy",
"dataset_id": "e2",
"st... | {
"score": {
"mean_loss": 4.865337643877865e-08,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 1.0
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 4.4305419777401767e-08,
"example_count": 300,
... |
303c8daf-643c-4fa1-ba9a-6ea50b02a828 | easy | khushidahi | 2026-08-03 02:59:27.063688+00:00 | succeeded | 503a1c711fefccb0059c139e49fe2779ba9953755fedc80b50659b1da46eda45 | 14,367 | 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": "303c8daf-643c-4fa1-ba9a-6ea50b02a828",
"created_at": "2026-08-03 02:59:27.063688+00:00",
"db_md5": "3b19d8801eebe3772518d2f8d8a7df10",
"submitter": "khushidahi",
"github_login": "khushidahi",
"run_id": "e90bc414-b61e-4e2a-94b9-87f9038054fb",
"tier": "easy",
"dataset_id": "e5",
"status": "succ... | {
"score": {
"mean_loss": 2.2034285068511963,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.01083333371207118
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.1981353759765625,
"example_count": 600... |
303dc869-1f4f-401a-bf71-60e94df95e49 | easy | AFeinfield | 2026-08-28 00:27:02.720567+00:00 | succeeded | a0ee421982bac33c0dfc402294e74848df92ef826fc996ca82c7eeba0fd2fd28 | 15,435 | null | """173: 161 + TIME-SCHEDULED X-SIZE CURRICULUM (comparator bootstrap).
Every variant plateaus at loss ~2.2: a gradient chicken-and-egg where
each operator needs the others correct to receive clean signal. The
carry op got its curriculum from the identity bias (158); this gives the
comparator one: small-x rows dominate... | {
"id": "303dc869-1f4f-401a-bf71-60e94df95e49",
"created_at": "2026-08-28 00:27:02.720567+00:00",
"db_md5": "d4a48cd7413c8971dfe067318ae07aa7",
"submitter": "AFeinfield",
"github_login": "AFeinfield",
"run_id": "49be2a84-03bd-42ca-a473-4bd20fcbafa9",
"tier": "easy",
"dataset_id": "e1",
"status": "succ... | {
"score": {
"mean_loss": 2.129945158958435,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0833333320915699
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.106865167617798,
"example_count": 100,
... |
304051f4-934f-46e8-bffb-c10a57fdde7b | easy | KaustubhKumar05 | 2026-08-29 04:54:44.729148+00:00 | succeeded | 89285038b28e448631ad6034fb5a7007c1cf5db04998c8d5583268facf41ab02 | 12,682 | null | """op2_t1w -- stock cell_op2 + T=1 loss upweighting (the scored rung).
Forked from cell_op2. ORIGINAL DOC BELOW.
cell_op2 -- ONE shared squaring operator with internal depth, applied T times.
WHY THIS SHAPE (verified on the real data, not assumed):
answer(T=t) == x squared t times mod N 20000 / 20000 e... | {
"id": "304051f4-934f-46e8-bffb-c10a57fdde7b",
"created_at": "2026-08-29 04:54:44.729148+00:00",
"db_md5": "de146efc427260a28a698b29f1f546c3",
"submitter": "koz",
"github_login": "KaustubhKumar05",
"run_id": "4544738f-0a16-4e4c-9701-d8bdf736d4aa",
"tier": "easy",
"dataset_id": "e5",
"status": "succee... | {
"score": {
"mean_loss": 2.495767884758138,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.012083333370586237
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.5092796252981966,
"example_count": 600... |
3042ba46-a918-440a-996c-556f625b2e1b | easy | AFeinfield | 2026-08-27 22:16:09.569738+00:00 | succeeded | 82ee9a809fa3260504112ca47ed47d24df8557f06ade4766f27ad749a6c78d2e | 12,086 | null | """155: 152 + digit-decode init scaled to the working range
(randn*3.0 — still random; the decode no longer has to grow 100x).
151 notes: throughput/learnability pass.
Dynamic alignments (m <= batch N-width), carry passes 6 post-square / 3
in reduction, single correction try, explicit y/10 feature (carry is
linear in... | {
"id": "3042ba46-a918-440a-996c-556f625b2e1b",
"created_at": "2026-08-27 22:16:09.569738+00:00",
"db_md5": "a3e2330961bb7e57846c9220189e9a1c",
"submitter": "AFeinfield",
"github_login": "AFeinfield",
"run_id": "ac8a0c74-af89-4bb0-9133-7c0f874c8068",
"tier": "easy",
"dataset_id": "e5",
"status": "succ... | {
"score": {
"mean_loss": 2.180303616948967,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.012500000037252903
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.180891880540035,
"example_count": 600,... |
304837e7-7c1f-4eba-9267-669fcd9dda14 | easy | viridale | 2026-08-05 01:07:29.131876+00:00 | succeeded | 3508dd9e8c61ed6ee03b7c115420d8092bbde2c8858543c63e01b955f59b5070 | 8,832 | null | """Clean hierarchical two-state recursive reasoner.
hypothesis: flat recurrence fails to separate fast latent work from slow answer-state
refinement; a coupled inner z / outer y recursion with final-string supervision at
completed outer cycles may learn a reusable iterative procedure.
axis: arch
target: offici... | {
"id": "304837e7-7c1f-4eba-9267-669fcd9dda14",
"created_at": "2026-08-05 01:07:29.131876+00:00",
"db_md5": "6791e3bcb9f2873f05aeac084cdc0161",
"submitter": "priormancer",
"github_login": "viridale",
"run_id": "da8e7fc6-ddcf-4b37-82cb-a165e90cd122",
"tier": "easy",
"dataset_id": "e1",
"status": "succe... | {
"score": {
"mean_loss": 2.1832228899002075,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0833333320915699
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.203216791152954,
"example_count": 100,
... |
30492df1-91bb-4f60-845f-6eef1c403381 | easy | yashkant | 2026-08-19 02:08:32.816527+00:00 | succeeded | 3ae1a37857aee4fed771a0b4c0660f750c47458fcd0c7b8bfc343eb7a30075ed | 45,397 | null | """Round2219 phase-balanced initialization for an anchored recurrent branch.
The confirmed correction-reset parent receives one width-8 recurrent path.
Its live transition and frozen initial anchor execute in parallel, so their
difference is exactly zero at construction and the parent function is kept.
The candidate s... | {
"id": "30492df1-91bb-4f60-845f-6eef1c403381",
"created_at": "2026-08-19 02:08:32.816527+00:00",
"db_md5": "0be9082b1b049610719f6dea1b3e55b4",
"submitter": "Yash Kant",
"github_login": "yashkant",
"run_id": "13977551-b0c4-4eff-926c-c21234af731b",
"tier": "easy",
"dataset_id": "e5",
"status": "succeed... | {
"score": {
"mean_loss": 2.148508135635394,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.010833333345750967
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.152998356541176,
"example_count": 600,... |
304ce55b-17dc-4f70-9c5c-a77717e9f648 | easy | TZZheng | 2026-08-06 21:06:07.624736+00:00 | succeeded | d63dc51ac327fb3105a6082a42efc5f68433cda9c21597a4e27c2547a3f6bb0f | 7,300 | null | """Self-contained standard Transformer scaling-sweep submission."""
from __future__ import annotations
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
assert_model_state,
)
VARIANT = {'block... | {
"id": "304ce55b-17dc-4f70-9c5c-a77717e9f648",
"created_at": "2026-08-06 21:06:07.624736+00:00",
"db_md5": "897811a993d1349f10f413b9cc1726fd",
"submitter": "TZZheng",
"github_login": "TZZheng",
"run_id": "d58d4afb-ef4b-4d6c-a6b4-602e6b240c97",
"tier": "easy",
"dataset_id": "e1",
"status": "succeeded"... | {
"score": {
"mean_loss": 2.192582845687866,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.07333333045244217
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.404803514480591,
"example_count": 100,
... |
304ceb1d-4f41-4226-901d-d2c4ea87df56 | easy | DDanlov | 2026-08-10 08:48:54.080875+00:00 | failed | 92eea01ad161f927da16140c4c5c925f2c2b6ead6149b2ed43f26a7259983cf7 | 16,397 | 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": "304ceb1d-4f41-4226-901d-d2c4ea87df56",
"created_at": "2026-08-10 08:48:54.080875+00:00",
"db_md5": "a60ee417a92c3d4130569185fe7bc14c",
"submitter": "DDanlov",
"github_login": "DDanlov",
"run_id": "0be9f045-08dc-4fa0-8ff1-4e267de8c1af",
"tier": "easy",
"dataset_id": "e5",
"status": "failed",
... | null
|
3050bccc-bdf7-4421-9eb0-b26891ec1273 | easy | viridale | 2026-08-03 21:37:39.973883+00:00 | succeeded | 4cda0ca962056cf33a776d62d228d1dc0ee5863efb6b881a441e94825634814b | 144,966 | null | """CUDA-wide mismatch-triggered broad evidence scan.
hypothesis: H100 fallback latency is dominated by Python/kernel launch overhead;
a CUDA-only 4194304-atom chunk uses measured memory headroom to reduce it.
axis: arch
target: revised h1 Max T at least1 with faster unsupported-route commitment.
expected_delta: pr... | {
"id": "3050bccc-bdf7-4421-9eb0-b26891ec1273",
"created_at": "2026-08-03 21:37:39.973883+00:00",
"db_md5": "990f3a734d7da60a1d7b7352bea851f2",
"submitter": "priormancer",
"github_login": "viridale",
"run_id": "3800c2d3-23ee-478e-8eee-585a780774dd",
"tier": "easy",
"dataset_id": "e3",
"status": "succe... | {
"score": {
"mean_loss": 0.0030203700887918954,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 1.0
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 0.0030203700470348266,
"example_count": 800,
... |
30572523-57a1-4230-8b86-00bc99318e87 | easy | sapient-sapiens | 2026-08-18 12:50:44.305257+00:00 | succeeded | a1385dfeac7ce5f085b7df39ce2b845570c41e86c2bbf5c62155a4832a8ed713 | 29,237 | null | """Depth-quantized recurrence with a mixture-over-depth objective.
The recurrent state is a bank of right-aligned digit slots. Every operator
application is followed by a soft quantization back onto the token simplex, and
the digit logits produced by that quantization are the answer logits at that
depth. There is no s... | {
"id": "30572523-57a1-4230-8b86-00bc99318e87",
"created_at": "2026-08-18 12:50:44.305257+00:00",
"db_md5": "f5f64960c51423621d139310e8b36a59",
"submitter": "Ertondy",
"github_login": "sapient-sapiens",
"run_id": "8461a9cf-5c6f-4b20-aa13-30d352494339",
"tier": "easy",
"dataset_id": "e9",
"status": "su... | {
"score": {
"mean_loss": 3.7963074445724487,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.02666666731238365
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 3.743250846862793,
"example_count": 90,
... |
3058b2b9-4656-45df-9463-298da3bc23ca | easy | sapient-sapiens | 2026-08-19 18:46:44.205625+00:00 | succeeded | 52fa23044dde388af9c351c0fba49483a7198959467bf4e812377532c138d4b7 | 33,710 | null | """Depth-quantized recurrence with a mixture-over-depth objective.
The recurrent state is a bank of right-aligned digit slots. Every operator
application is followed by a soft quantization back onto the token simplex, and
the digit logits produced by that quantization are the answer logits at that
depth. There is no s... | {
"id": "3058b2b9-4656-45df-9463-298da3bc23ca",
"created_at": "2026-08-19 18:46:44.205625+00:00",
"db_md5": "39f8571338c31ba5eed0216457e6f1af",
"submitter": "Ertondy",
"github_login": "sapient-sapiens",
"run_id": "0661533c-24ca-42f8-81eb-e97a170de94a",
"tier": "easy",
"dataset_id": "e5",
"status": "su... | {
"score": {
"mean_loss": 2.2007967464807034,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0037500000124176345
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.19556889325514,
"example_count": 600... |
305d4b26-6d0a-4f12-a295-712914d363fe | easy | erdavis0 | 2026-08-21 11:32:11.220433+00:00 | succeeded | aaf896656a39960006176593f51c42466ac987207d80a5db9c8cd173ee6d2562 | 9,256 | null | """Context-compiled monotone executor with sealed tied recurrence."""
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,
)
WI... | {
"id": "305d4b26-6d0a-4f12-a295-712914d363fe",
"created_at": "2026-08-21 11:32:11.220433+00:00",
"db_md5": "92ddd807eb4760c5cd1a4c89a3bb7a3e",
"submitter": "Ethan",
"github_login": "erdavis0",
"run_id": "e0ddd778-ec01-4bd5-a880-a1f300c7ff85",
"tier": "easy",
"dataset_id": "e6",
"status": "succeeded",... | {
"score": {
"mean_loss": 2.2302401065826416,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0335585605353117
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.1535136699676514,
"example_count": 60,
... |
305eb0a4-3d0e-422e-949c-0a5959ede835 | easy | isaac0804 | 2026-08-06 22:25:01.837503+00:00 | succeeded | a60a70156991e5ac7e5205255e1a9fcec6fb60d7311e56b0cb67f9f7a7bd1fe7 | 9,956 | null | """E3 candidate: hard N-routing MoE + non-gradient "loss-free" expert-load
balancing -- a targeted revisit of submissions/e3_hard_routing's specific
failure, not a blind new mechanism.
Context (docs/findings.md): e3_hard_routing (straight-through hard routing
over NUM_EXPERTS=4 shared LoRA deltas, gated by N) made the... | {
"id": "305eb0a4-3d0e-422e-949c-0a5959ede835",
"created_at": "2026-08-06 22:25:01.837503+00:00",
"db_md5": "5c7fdbc99cccf5f67aaaeaa0df36cce8",
"submitter": "Isaac Yong",
"github_login": "isaac0804",
"run_id": "d0392dad-c83f-4982-880f-2ef77fc02ce8",
"tier": "easy",
"dataset_id": "e3",
"status": "succe... | {
"score": {
"mean_loss": 11.617262030986765,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.007500000009313225
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 11.663856559307273,
"example_count": 80... |
30624ffa-2091-4b40-9f17-a4805e0dc7ae | easy | DDanlov | 2026-08-29 14:28:09.878355+00:00 | succeeded | 4d16fb8952769b6d66d93ad1bcf04fc8a45bc7f6cc1c889be02c131acbe35c9f | 27,037 | null | """
trial_l08_l3_d768_k48_rel005.py
L08: L=3, D=768, K_max=48, rel_force=0.5% (0.005), progress_thresh=0.75, lr=0.0035
"""
import math
import time
import contextlib
from dataclasses import dataclass
from typing import Optional, Tuple, Dict, Any, List, Union
import torch
import torch.nn as nn
import torch.nn.functional... | {
"id": "30624ffa-2091-4b40-9f17-a4805e0dc7ae",
"created_at": "2026-08-29 14:28:09.878355+00:00",
"db_md5": "9e7c59c49929557d475d041f7f8cba4d",
"submitter": "DDanlov",
"github_login": "DDanlov",
"run_id": "ad990d96-3daf-46f0-963f-3c35d3d77202",
"tier": "easy",
"dataset_id": "e5",
"status": "succeeded"... | {
"score": {
"mean_loss": 2.298406591436878,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.014583333345750966
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.3250923234251046,
"example_count": 600... |
3068f1c1-4a27-45f5-898a-b2df5dbd660c | easy | falloficaruss | 2026-08-06 05:32:48.680690+00:00 | succeeded | a12208a58faf462e8b49c8ede2f425d7a5093e2dff25000348e1dd1383606702 | 10,463 | null | """Looped encoder v4 — generalize, don't memorize.
Evidence so far
---------------
m1: digit-prior collapse, score ≈ 0.07%, Max T none.
e1 d54f8d8f (train@8/eval@24): train 95% exact, test 5% — depth mismatch.
e1 70f45c59 (matched depth 12, batch 512): train 97%, test 1.3%, 238 steps/60s
— pure memorization of the 6... | {
"id": "3068f1c1-4a27-45f5-898a-b2df5dbd660c",
"created_at": "2026-08-06 05:32:48.680690+00:00",
"db_md5": "a6343bb0c851bc57e3e86b21220df86b",
"submitter": "Abhishek Shinde",
"github_login": "falloficaruss",
"run_id": "d67e109d-c8cb-4b43-9a94-4b45737202eb",
"tier": "easy",
"dataset_id": "e3",
"status... | {
"score": {
"mean_loss": 2.280895767489394,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.00625
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.3130359340727105,
"example_count": 800,
"... |
3071e574-6ac6-4a9f-b901-982059e7d8cc | easy | isaac0804 | 2026-08-07 16:14:10.622007+00:00 | succeeded | 4a03526d77634514923b6c9053dc84cd558c037c110829a28a590605676ce4cf | 5,355 | null | """Fixed-depth loop: the shared block is applied a constant number of times
(FIXED_ITERS) to every row, unconditionally -- no T-parsing, no per-row
`torch.where` masking. Depth is structure, not input-dependent routing. On
E3 (all T=2) this means the block runs 2 surplus iterations beyond the
answer, which the weights ... | {
"id": "3071e574-6ac6-4a9f-b901-982059e7d8cc",
"created_at": "2026-08-07 16:14:10.622007+00:00",
"db_md5": "cfe8ac13a1143ed369706f14163cca1f",
"submitter": "Isaac Yong",
"github_login": "isaac0804",
"run_id": "fcdd0c32-2137-42b5-b7b4-993e85f06994",
"tier": "easy",
"dataset_id": "e3",
"status": "succe... | {
"score": {
"mean_loss": 5.6306270937108565,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.005000000004656613
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 5.560270076345685,
"example_count": 800... |
30731384-e6cb-4978-a08e-524cdc9eee55 | easy | karanganesan | 2026-08-05 10:04:54.981018+00:00 | succeeded | 7af54e899459a26200a71b0c876d5e3d942c2eb7e1c418be459891c847072686 | 22,081 | null | """Parametric looped-transformer family (P1).
One weight-tied transformer block applied k times in latent space. Config
flags cover four P1 families with one file:
looped recall=0 gated=0 tfilm=0 plain weight-tied loop
looped-recall recall=1 re-inject the input embedding each
... | {
"id": "30731384-e6cb-4978-a08e-524cdc9eee55",
"created_at": "2026-08-05 10:04:54.981018+00:00",
"db_md5": "255e2f927dd274a2579557bd8a91931d",
"submitter": "Karan Ganesan",
"github_login": "karanganesan",
"run_id": "833b6776-efe9-4706-9afc-c61d2ebbc37d",
"tier": "easy",
"dataset_id": "e1",
"status": ... | {
"score": {
"mean_loss": 8.90067434310913,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.03499999921768904
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 8.726667404174805,
"example_count": 100,
... |
3076173b-d5c6-43cc-b2f5-59cbc398f483 | easy | velocizapkar | 2026-08-06 22:32:16.481124+00:00 | succeeded | 634ceb3a84105aaefe6d8b9f8ccbbb18ac5bb1fa0ae0e86edd94f45e4ad23a23 | 55,676 | null | """Fast packed residual-free Attention-Residual transformer with Muon.
The model is a generic learned sequence model. It contains no parser,
arithmetic primitive, task constant, generated data, label feedback, or answer
construction. Every macro-state is produced by a serial pack of attention and
MLP transformations... | {
"id": "3076173b-d5c6-43cc-b2f5-59cbc398f483",
"created_at": "2026-08-06 22:32:16.481124+00:00",
"db_md5": "9546c636816918eeffb9b712a0ef6931",
"submitter": "Aakanksh Zarapkar",
"github_login": "velocizapkar",
"run_id": "0a2a1f9f-1acd-4a9c-bdcc-7d63d08e5ddb",
"tier": "easy",
"dataset_id": "e3",
"statu... | {
"score": {
"mean_loss": 3.2204928398132324,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.015624999534338713
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 3.1965532302856445,
"example_count": 80... |
3088550a-3793-4cb5-8a04-9b9a6a54923e | easy | jayyala | 2026-08-16 22:31:26.788786+00:00 | succeeded | b5acf766cf76f2016093f66fe292fb0198f4803d7ccb9de5f3a468aec4dbb9ce | 13,296 | null | """A scalable recurrent machine with canonical categorical byte states.
Each integer is represented by a short little-endian tape of 256-way slots.
One tied bidirectional mixing cell learns the unknown state transition from the
current tape and the prompt context, then its categorical output is fed back
for exactly T ... | {
"id": "3088550a-3793-4cb5-8a04-9b9a6a54923e",
"created_at": "2026-08-16 22:31:26.788786+00:00",
"db_md5": "10dc525bee82f997424fae858ca3f88b",
"submitter": "jayyala",
"github_login": "jayyala",
"run_id": "a41151aa-ef15-4a53-ab3e-e332a1caeed1",
"tier": "easy",
"dataset_id": "e1",
"status": "succeeded"... | {
"score": {
"mean_loss": 5.011166155338287,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.8583333194255829
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 1.0447760820388794,
"example_count": 100,
... |
308cdca1-35eb-4733-aadb-3f1bcd276fd3 | easy | chad-atexpedient | 2026-08-04 21:49:36.705771+00:00 | succeeded | 4fe1c1c08675c210d3f66a725dd37f7833fe74a4b911cda25867b0edba2b54d4 | 4,933 | null | """R016: one-pass RoPE Transformer with a lightweight local digit mixer."""
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 = 1... | {
"id": "308cdca1-35eb-4733-aadb-3f1bcd276fd3",
"created_at": "2026-08-04 21:49:36.705771+00:00",
"db_md5": "41156dc8655d29ae3d6119f1cff29d86",
"submitter": "chad-atexpedient",
"github_login": "chad-atexpedient",
"run_id": "1cd60dc0-9c5e-4479-bd20-137185c1e3ad",
"tier": "easy",
"dataset_id": "e5",
"st... | {
"score": {
"mean_loss": 4.072245017250161,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.004166666666666667
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 4.748682982183893,
"example_count": 600,... |
3095400d-bba0-45be-9bd8-6a26674ca7f4 | easy | khushidahi | 2026-08-03 02:57:31.987229+00:00 | succeeded | b7446a145df32470dd8870685f27928c70b8cc2893636ed5d4258d0e3ab32bc2 | 14,365 | null | """R6 structured decimal-workspace model for One Layer Deeper.
The model does not implement multiplication, modular reduction, or the public
recurrence. It uses the public tokenizer structure to build learned, right-
aligned decimal tapes for N, X, and T, then applies a shared neural transition
to a mutable work tape.... | {
"id": "3095400d-bba0-45be-9bd8-6a26674ca7f4",
"created_at": "2026-08-03 02:57:31.987229+00:00",
"db_md5": "3aa3693f8d86aaee9466ccbc65636fc0",
"submitter": "khushidahi",
"github_login": "khushidahi",
"run_id": "9537e35d-4d47-40e7-920f-b250b9348692",
"tier": "easy",
"dataset_id": "e5",
"status": "succ... | {
"score": {
"mean_loss": 2.212110161781311,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.008333333767950535
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.209249258041382,
"example_count": 600,... |
309c97d9-f2b0-47fe-b43d-4927d6a6bcc6 | easy | gauravmishra | 2026-08-10 03:42:05.600442+00:00 | succeeded | 440e453d70006f5f1b70b9795979c6a2dff940fe9e707d1c1ea52f9e75051fa6 | 32,486 | null | """Information-gain candidate ig60_e25_trm_lite_e5 (e18_tie_fallback)."""
from __future__ import annotations
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
TokenLossBatch,
assert_model_st... | {
"id": "309c97d9-f2b0-47fe-b43d-4927d6a6bcc6",
"created_at": "2026-08-10 03:42:05.600442+00:00",
"db_md5": "07288d1482457fc09ad98d35f5a0994c",
"submitter": "Gaurav Mishra",
"github_login": "gauravmishra",
"run_id": "2c36dd0f-ec19-4c62-a378-bd658dfea169",
"tier": "easy",
"dataset_id": "e5",
"status": ... | {
"score": {
"mean_loss": 2.2354109751316678,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.005000000012417635
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.242991213306718,
"example_count": 600... |
309e563f-ba27-4667-b45e-76b6ee3e5f8c | easy | yashkant | 2026-08-09 08:03:20.090330+00:00 | succeeded | bf8c051fc6405dc7a1e262f63a8970572b2f3c544f2e415574bd7ae0ce9204ca | 23,143 | null | """Round 390 pre-block1 TRM with a ceil(log2(T))+1 call schedule, capped at 4."""
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,
)
fr... | {
"id": "309e563f-ba27-4667-b45e-76b6ee3e5f8c",
"created_at": "2026-08-09 08:03:20.090330+00:00",
"db_md5": "caa96ac24b6b3383ee8153bdd06582c0",
"submitter": "Yash Kant",
"github_login": "yashkant",
"run_id": "f4faa8d9-7127-46ba-9524-4b6c1803a8de",
"tier": "easy",
"dataset_id": "e2",
"status": "succeed... | {
"score": {
"mean_loss": 3.9139384627342224,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.3135416842997074
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 6.3523173332214355,
"example_count": 300,... |
309eb8d3-fbee-49ea-bca4-0031200c840f | easy | gauravmishra | 2026-08-07 19:03:12.011916+00:00 | succeeded | 9140fef522ec6a9afeb558e69343f534c498c0164ab7648853f9bbf90a742043 | 3,726 | null | """Four-pass Transformer with four independently parameterized blocks."""
from __future__ import annotations
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
assert_model_state,
)
D_MODEL = 1... | {
"id": "309eb8d3-fbee-49ea-bca4-0031200c840f",
"created_at": "2026-08-07 19:03:12.011916+00:00",
"db_md5": "067dd19f95b3a2b025d36561c727f6b0",
"submitter": "Gaurav Mishra",
"github_login": "gauravmishra",
"run_id": "e6f366a8-1072-4de9-9b49-071bbbfde8f4",
"tier": "easy",
"dataset_id": "e1",
"status": ... | {
"score": {
"mean_loss": 2.6118286848068237,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.004999999888241291
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.8294677734375,
"example_count": 100,
... |
30a0be28-a637-42a7-b7e2-1fb139db33d7 | easy | yashkant | 2026-08-15 11:59:48.415414+00:00 | succeeded | 1bae221b965dc02164b760f423fd9f13a432696042ed6e215f8f79e29e7e7834 | 44,981 | null | """Round1935 D273 four-digit-modulus residual expert.
The exact Round1917 D273 sparse-SAM8 carrier remains intact. A small bounded
residual is appended after all parent tensors and activated only when the
public N field contains at least four decimal digits. The candidate derives
that residual from the immutable N-f... | {
"id": "30a0be28-a637-42a7-b7e2-1fb139db33d7",
"created_at": "2026-08-15 11:59:48.415414+00:00",
"db_md5": "8f44323ab5b06ea9d57c404ddf246080",
"submitter": "Yash Kant",
"github_login": "yashkant",
"run_id": "05d33070-346f-40d2-af21-2d18ae373e04",
"tier": "easy",
"dataset_id": "e4",
"status": "succeed... | {
"score": {
"mean_loss": 2.069790297634504,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0104166666790843
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.0727036590889787,
"example_count": 1200,... |
30a0efa4-b74e-4c5a-9d03-f7a65a6eb11c | easy | lzy54 | 2026-08-29 08:36:35.133105+00:00 | succeeded | f967c3bfea2806a17090394364d66f815812d86d38acc90cb7b0b8126f881c01 | 4,539 | null | """Phase 2 C2: shared block x4 with residual concat recall."""
from __future__ import annotations
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
assert_model_state,
)
D_MODEL = 128
NUM_HEAD... | {
"id": "30a0efa4-b74e-4c5a-9d03-f7a65a6eb11c",
"created_at": "2026-08-29 08:36:35.133105+00:00",
"db_md5": "649789da2b5f7837ff738282292ec92c",
"submitter": "lzy54",
"github_login": "lzy54",
"run_id": "a61ee0d5-f32e-47fd-8245-aef35527d72a",
"tier": "easy",
"dataset_id": "e8",
"status": "succeeded",
... | {
"score": {
"mean_loss": 2.7187843322753906,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.03279857523739338
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.6569736003875732,
"example_count": 85,... |
30a28eb5-fff2-4700-85c7-766c21057438 | easy | DDanlov | 2026-08-23 21:51:57.088483+00:00 | succeeded | 3071b91085f18afbea5f4408079dfe8127a028259638408d110b4b2b470adb98 | 31,834 | null | """
Unified Hybrid Masked Deliberation Model (71.9M Parameters) - Tier 1 (Pure Skipless, No Cross-Attention)
Architecture:
- Scale: dim=2048, num_heads=32, d_ff=8192 (Exactly 71,889,920 parameters)
- Hybrid Prefix-Bidirectional + Causal Target Attention Mask
- Token Layout: [Prompt (L_p), 4 Pause Registers, Target Toke... | {
"id": "30a28eb5-fff2-4700-85c7-766c21057438",
"created_at": "2026-08-23 21:51:57.088483+00:00",
"db_md5": "5db0adc2ab754a192a7886c4976d4aa1",
"submitter": "DDanlov",
"github_login": "DDanlov",
"run_id": "0dc02be3-833f-4a5f-bb8d-5d94e41959a1",
"tier": "easy",
"dataset_id": "e5",
"status": "succeeded"... | {
"score": {
"mean_loss": 3.3066962502537995,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 3.302940074638401,
"example_count": 600,
"exac... |
30b2006d-c9b8-4e8b-bb40-b666e336b566 | easy | shirvani-jr | 2026-08-28 10:53:25.489920+00:00 | succeeded | 6e92bc87c307530e53faf008feedb887887e14837e82bbcb80a664ed8d6dfc71 | 51,183 | 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": "30b2006d-c9b8-4e8b-bb40-b666e336b566",
"created_at": "2026-08-28 10:53:25.489920+00:00",
"db_md5": "ea8cfb8774e478067705077970765c2a",
"submitter": "Ali",
"github_login": "shirvani-jr",
"run_id": "246d9710-29d0-4de6-9b44-148af036eca4",
"tier": "easy",
"dataset_id": "e1",
"status": "succeeded"... | {
"score": {
"mean_loss": 2.084315598011017,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.08333333395421505
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.19461989402771,
"example_count": 100,
... |
30b36006-05b8-4939-93da-5b19cba2f616 | easy | DDanlov | 2026-08-15 16:06:46.297648+00:00 | succeeded | 67846bc4e9dc14db5c35c9f4fd6288ae2670cd78050c8c3de72fab700a455904 | 18,369 | null | """
Official Submission for One Layer Deeper Challenge
Architecture: Model A - Scaled Symplectic / Hamiltonian Invariant Flow Model (HamFlowNet)
- Canonical Symplectic Phase Rotations + Orthogonal Channel Mixing
- dim=1024, d_hidden=2048, Pure No-Norm No-Skip DKS, RohanShampoo with lambda=0.12
"""
from __future__ impo... | {
"id": "30b36006-05b8-4939-93da-5b19cba2f616",
"created_at": "2026-08-15 16:06:46.297648+00:00",
"db_md5": "d8b715ff4cbe7c702858bfc297e18d54",
"submitter": "DDanlov",
"github_login": "DDanlov",
"run_id": "8a38118a-028e-49d5-abc6-4f97696df163",
"tier": "easy",
"dataset_id": "e5",
"status": "succeeded"... | {
"score": {
"mean_loss": 15.864406718263005,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.00625
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 26.12913387880197,
"example_count": 600,
"... |
30b6380b-ff0f-4a08-8024-dab6f696f836 | easy | DDanlov | 2026-08-07 09:27:47.591203+00:00 | succeeded | 894a8adbb4977778772528a7fb82659fe13d071732e7c10dac66517f90d8e7b3 | 11,883 | null | """
Ultra-Scaled Official Competition Submission:
Architecture Specs:
- Depth: 12 Transformer Decoder Layers (NUM_LAYERS = 12)
- Width: D_MODEL = 1024
- Attention Heads: 128 Heads (NUM_HEADS = 128)
- Recurrent Unrolled Steps: K = 120 PLER Steps (NUM_STEPS = 120)
- Optimizer: RohanShampoo & BlockShampoo (Full Inter-Laye... | {
"id": "30b6380b-ff0f-4a08-8024-dab6f696f836",
"created_at": "2026-08-07 09:27:47.591203+00:00",
"db_md5": "78f82516286c05ff746615793799f4b5",
"submitter": "DDanlov",
"github_login": "DDanlov",
"run_id": "ac243741-2d3c-4839-a01d-e666de941683",
"tier": "easy",
"dataset_id": "e1",
"status": "succeeded"... | {
"score": {
"mean_loss": 3.204512357711792,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.023333333432674408
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 3.1202056407928467,
"example_count": 100... |
30b8a3ba-07ff-4d3f-b4a1-df0efe65436b | easy | Yalyenea | 2026-08-12 12:57:16.703641+00:00 | succeeded | 2e372fc533302a76150be75cdeffdbdf6f15077674bf40ae208029f8710b9f29 | 11,729 | null | """Digit-class gated recurrence with 20% residual dropout."""
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,
as... | {
"id": "30b8a3ba-07ff-4d3f-b4a1-df0efe65436b",
"created_at": "2026-08-12 12:57:16.703641+00:00",
"db_md5": "4cd75047e7a6b3f494b576ab0f1310b5",
"submitter": "yfff",
"github_login": "Yalyenea",
"run_id": "151c3027-a812-4510-8e6f-a0d15c99de37",
"tier": "easy",
"dataset_id": "e5",
"status": "succeeded",
... | {
"score": {
"mean_loss": 3.3056719122999407,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.012916666703919569
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.2818761278161017,
"example_count": 60... |
30bba4aa-e304-4019-9394-c528ae31ddbd | easy | yashkant | 2026-08-10 21:18:21.400501+00:00 | succeeded | 6765edee8ea449548c89c857007b55712e4790a8183afb82b951b42e7ad3e5f7 | 24,350 | null | """Round 453 stratum-balanced CE with 1% warmup on the cap-4 transition."""
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 tor... | {
"id": "30bba4aa-e304-4019-9394-c528ae31ddbd",
"created_at": "2026-08-10 21:18:21.400501+00:00",
"db_md5": "2f0815a2fe91d71ceaabfd6b4a2cbeed",
"submitter": "Yash Kant",
"github_login": "yashkant",
"run_id": "5f9b1dba-05e3-45ef-86fe-14d16eddcfa5",
"tier": "easy",
"dataset_id": "e5",
"status": "succeed... | {
"score": {
"mean_loss": 2.2061066307322825,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.010833333345750967
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.2237609031489076,
"example_count": 60... |
30c07093-7aca-4664-94e9-f9553e92b262 | easy | ehonig | 2026-08-21 01:48:23.006441+00:00 | succeeded | a8101c7bf20f5fb52423225be6359acd59e36c1d9a8a08942c855acf40ae70ac | 31,882 | null | """A plain Transformer for repeated modular squaring, and nothing else yet.
This is a deliberate restart. The previous submission accumulated a field
parser, a carry scan, a learned reciprocal, periodic features, a digit
bottleneck, a depth selector and two cell types, most of them tuned around a
memorisation table th... | {
"id": "30c07093-7aca-4664-94e9-f9553e92b262",
"created_at": "2026-08-21 01:48:23.006441+00:00",
"db_md5": "75d8fa2efbdc026e18b4eec8692e50b8",
"submitter": "Edouardo Honig",
"github_login": "ehonig",
"run_id": "ef03e903-8687-455c-bd13-d13701515275",
"tier": "easy",
"dataset_id": "e4",
"status": "succ... | {
"score": {
"mean_loss": 2.278085926086174,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.00739583333566164
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.4493729026387685,
"example_count": 1200... |
30c22104-efbb-4dfa-8c22-dcec8dece162 | easy | jordanrubin | 2026-08-13 18:04:41.402001+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": "30c22104-efbb-4dfa-8c22-dcec8dece162",
"created_at": "2026-08-13 18:04:41.402001+00:00",
"db_md5": "3f710550312ea469d13b008d83980242",
"submitter": "Jordan Rubin",
"github_login": "jordanrubin",
"run_id": "c8651c60-5488-4ab7-82ae-ccc975411360",
"tier": "easy",
"dataset_id": "e5",
"status": "s... | {
"score": {
"mean_loss": 2.8713181356928823,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.023333332889402907
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.228764378837406,
"example_count": 600... |
30c2fffb-c425-4350-940e-d2eeb1f5a476 | easy | DDanlov | 2026-08-14 07:33:04.998505+00:00 | succeeded | 681e409b7573146f8da5882dd37edd6ac3f6b07889fe73d34dd8e60d7d0ad724 | 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": "30c2fffb-c425-4350-940e-d2eeb1f5a476",
"created_at": "2026-08-14 07:33:04.998505+00:00",
"db_md5": "1fd8b837bee0e9ee7bf1e9b85bcc63d8",
"submitter": "DDanlov",
"github_login": "DDanlov",
"run_id": "44a45eb0-154c-4a74-b490-e077206f37e8",
"tier": "easy",
"dataset_id": "e3",
"status": "succeeded"... | {
"score": {
"mean_loss": 2.3976615785039233,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.018125000000000002
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.3970079838251634,
"example_count": 80... |
30c46054-7359-4671-aca9-7e817d348466 | easy | liam-gb | 2026-08-11 17:00:53.122434+00:00 | succeeded | a91c83a2dfd13a5dc896db83beb80c57739bf9970ca116d6f8827525f5a45060 | 13,278 | null | """rns_pm_gather_ct: rns_pm_gather + counted halting on a weight-tied
loop.
Changes vs rns_pm_gather: the sigmoid T-gates and per-step embeddings
are DELETED (per-iteration parameters past the trained depth receive no
gradient and are exactly the measured extrapolation gap). One shared
transition is applied hard-count... | {
"id": "30c46054-7359-4671-aca9-7e817d348466",
"created_at": "2026-08-11 17:00:53.122434+00:00",
"db_md5": "6905ef43ae1bcceded1feba840f41533",
"submitter": "liam-gb",
"github_login": "liam-gb",
"run_id": "db5901b1-57d1-4b76-9aad-0434000fd901",
"tier": "easy",
"dataset_id": "e1",
"status": "succeeded"... | {
"score": {
"mean_loss": 5.001043438911438,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.5050000175833702
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 8.787213325500488,
"example_count": 100,
... |
30c736d4-3b06-4e28-ad7d-3f489b790623 | easy | liam-gb | 2026-08-14 02:14:10.080885+00:00 | succeeded | 3c39c52b433ddd2675e5ce63fea469f9c9bd5e09a62e7e6fda5ccefcd8242585 | 14,760 | null | """rns_pm_gather_ct_widet_nore: widet with batch reuse off.
The wide readout is what makes long answers reachable; the trim is what pays
for it, by not running loop iterations whose update is multiplied by zero.
Base docstring:
Single change vs rns_pm_gather_ct: the readout carries one
distance-from-units slot per s... | {
"id": "30c736d4-3b06-4e28-ad7d-3f489b790623",
"created_at": "2026-08-14 02:14:10.080885+00:00",
"db_md5": "940211fbc28f5e5b59bd7c1149be5b02",
"submitter": "liam-gb",
"github_login": "liam-gb",
"run_id": "e827b254-130c-4da6-b91c-6e195df3f7b2",
"tier": "easy",
"dataset_id": "e5",
"status": "succeeded"... | {
"score": {
"mean_loss": 3.596735572828366,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.028333333327124517
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 4.204110145568848,
"example_count": 600,... |
30c90d31-b5ec-4b9a-a248-7b0bf05debf2 | easy | anirudh-chakravarthy | 2026-08-12 04:23:35.320463+00:00 | succeeded | 0ddffef4c93a7475e771ba422ae52bd579d12a63f58bcac0cb0074e23e3837ce | 15,456 | null | """A learned state transition applied once per requested time step."""
from __future__ import annotations
import math
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
TokenLossBatch,
asser... | {
"id": "30c90d31-b5ec-4b9a-a248-7b0bf05debf2",
"created_at": "2026-08-12 04:23:35.320463+00:00",
"db_md5": "81a68fd73353928aeb9d8c3ccdec906b",
"submitter": "Anirudh S Chakravarthy",
"github_login": "anirudh-chakravarthy",
"run_id": "2c9949a2-89e4-4ffb-95d5-8430a5903841",
"tier": "easy",
"dataset_id": "... | {
"score": {
"mean_loss": 2.9534463380044547,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.012500000012417634
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 3.3902143533988918,
"example_count": 60... |
30cdb468-6e90-45d4-83e1-e680bb6d7d5d | easy | DDanlov | 2026-08-13 22:43:19.792736+00:00 | failed | f7e404d838107f8d3c299291eef901aa43dda5f401ea0feda4fe7ec5cf5fb2cb | 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": "30cdb468-6e90-45d4-83e1-e680bb6d7d5d",
"created_at": "2026-08-13 22:43:19.792736+00:00",
"db_md5": "49855c75825d5c11432fd619c223b54f",
"submitter": "DDanlov",
"github_login": "DDanlov",
"run_id": "0933d332-d356-4691-90a4-14f1561f7262",
"tier": "easy",
"dataset_id": "e3",
"status": "failed",
... | null
|
30d0deb7-366b-4543-a585-a8724c0a4218 | easy | shreyash-chonkie | 2026-08-26 14:31:19.045461+00:00 | succeeded | 4a4a645b8f51fef10c2d7198110a63971802e66ef5b1dcb4068c572429ffe2cf | 17,635 | null | """Online hard-controller interpreter with executor-owned continuous memory."""
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_sta... | {
"id": "30d0deb7-366b-4543-a585-a8724c0a4218",
"created_at": "2026-08-26 14:31:19.045461+00:00",
"db_md5": "dfbe7b24fcd0ddb1f6c1f56728d78685",
"submitter": "Shreyash",
"github_login": "shreyash-chonkie",
"run_id": "6053f1dc-9256-4f29-b5c3-17de28f03404",
"tier": "easy",
"dataset_id": "e1",
"status": "... | {
"score": {
"mean_loss": 1.8665454983711243,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.06499999761581421
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 1.8040351867675781,
"example_count": 100... |
30d58587-c517-4ccb-be15-9a5ba0c2c6b1 | easy | khushidahi | 2026-08-01 21:33:33.631280+00:00 | succeeded | ec59bb6a0920283ef46bedc73748389ab7bc6bffcf2faa7902751be3e4156c16 | 4,045 | 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": "30d58587-c517-4ccb-be15-9a5ba0c2c6b1",
"created_at": "2026-08-01 21:33:33.631280+00:00",
"db_md5": "294da15a5433dc15d3fe91202e22e405",
"submitter": "khushidahi",
"github_login": "khushidahi",
"run_id": "16e11483-50e0-48c1-9faa-00e2a3d5ebb9",
"tier": "easy",
"dataset_id": "e5",
"status": "succ... | {
"score": {
"mean_loss": 3.0338011398843454,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.00541666670391957
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 3.0524850458307653,
"example_count": 600... |
30e5d4d9-5749-4b10-9f9d-35a3f9b4cebe | easy | himalalps | 2026-08-07 01:28:38.288397+00:00 | succeeded | 936be2755305fda3dc9760859cd1191a14de1abfce22f1359a58c947694bc260 | 64,891 | 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": "30e5d4d9-5749-4b10-9f9d-35a3f9b4cebe",
"created_at": "2026-08-07 01:28:38.288397+00:00",
"db_md5": "eba0403516d182046f5d05ee8d37763e",
"submitter": "Haoyu Tang",
"github_login": "himalalps",
"run_id": "75dde1bd-a35d-4472-8882-3d2953da9df8",
"tier": "easy",
"dataset_id": "e5",
"status": "succe... | {
"score": {
"mean_loss": 2.2126786387986965,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.004583333353511989
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.2181766033172607,
"example_count": 60... |
30e8809a-7cfd-475b-a723-2f755e5f8643 | easy | adambrzosko | 2026-08-30 16:52:49.170694+00:00 | succeeded | 925ee74a0a6050d4f28046c4c8a0658823284eeff9194daaf42ba2326ccf7824 | 14,139 | null | """One Layer Deeper — digit-slot recurrent machine ("dm").
The prompt is [N] digits(N) [X] digits(x) [T] digits(T) and the label is the
decimal expansion of x^(2^T) mod N, tail-aligned: the last prompt position
carries the units digit, the one before it the tens digit, and so on.
Architecture. The model parses the ... | {
"id": "30e8809a-7cfd-475b-a723-2f755e5f8643",
"created_at": "2026-08-30 16:52:49.170694+00:00",
"db_md5": "69c2afec6ae39d06a1b0a61c0f12d146",
"submitter": "Adam Brzosko",
"github_login": "adambrzosko",
"run_id": "8057f3de-10c3-497e-9f46-51554c5c9aa0",
"tier": "easy",
"dataset_id": "e5",
"status": "s... | {
"score": {
"mean_loss": 3.2583475113542617,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0066666666915019355
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 3.1829289962358005,
"example_count": 6... |
30eebfa7-a1d8-4310-b1ac-0e07fe370317 | easy | DESU-CLUB | 2026-08-30 19:50:33.231071+00:00 | succeeded | 512145c1f969ba1a12f4f69ba1d8af04597239399c6295ad343d36f5c752aa3b | 16,039 | null | """Discrete-state loop over a place-value tape, trained with orthogonalize-only Muon.
The recurrent state carried between squarings is the *digit string*, not a
residual stream. Each outer iteration reads (N, x) off an LSB-first place-value
tape, emits digit logits for one squaring, hard-argmaxes them (straight-throu... | {
"id": "30eebfa7-a1d8-4310-b1ac-0e07fe370317",
"created_at": "2026-08-30 19:50:33.231071+00:00",
"db_md5": "fe24752b5b6d14ae8fc4176d0aeb273a",
"submitter": "Low Keng Hoong, Warren",
"github_login": "DESU-CLUB",
"run_id": "8f0eeff8-e394-49a3-8415-494e3a1af9e1",
"tier": "easy",
"dataset_id": "e6",
"sta... | {
"score": {
"mean_loss": 1.211648061667403,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.6108108219262716
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 1.0681793689727783,
"example_count": 60,
... |
30f3c3cd-94bd-49ea-a67a-d7da9571bccf | easy | vatsalnar123 | 2026-08-18 03:02:08.301208+00:00 | succeeded | 081788572218c2eac52ee61acc9a8ff9bf1a115393ab9aeb937f763674636088 | 15,427 | null | """Weight-tied looped transformer where loop depth is the only T channel.
v3 versus v2: the T-digit token embeddings are replaced with a learned
placeholder before injection, so the prompt content carries no information
about T — the loop count (T + base) is the only way T can influence the
output. Rows sharing an x b... | {
"id": "30f3c3cd-94bd-49ea-a67a-d7da9571bccf",
"created_at": "2026-08-18 03:02:08.301208+00:00",
"db_md5": "80f9f237cf52aeb4990a3c117f0f4685",
"submitter": "vatsalnar123",
"github_login": "vatsalnar123",
"run_id": "c12fe2cb-5936-45bc-978e-863613e2c3b5",
"tier": "easy",
"dataset_id": "e3",
"status": "... | {
"score": {
"mean_loss": 3.582011791323604,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.015000000000000001
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 3.7499876376406003,
"example_count": 800... |
30f798ce-b615-4259-ac1f-cb9b5116347f | easy | sapient-sapiens | 2026-08-23 20:39:43.789015+00:00 | failed | 50281034562e83fd1f92b97c13250caf44700b72f462b244927f46841db34f77 | 18,319 | null | """42M Q-gated SwiGLU recurrent Transformer for the E5 campaign.
The arithmetic stream uses categorical token embeddings plus marker-derived
field and left-relative embeddings. N, X, and T are never decoded as numbers.
T-field values are excluded entirely from the arithmetic and halting streams.
All recurrent readout... | {
"id": "30f798ce-b615-4259-ac1f-cb9b5116347f",
"created_at": "2026-08-23 20:39:43.789015+00:00",
"db_md5": "2e7b9283992424a604293b4b34d67b61",
"submitter": "Ertondy",
"github_login": "sapient-sapiens",
"run_id": "01452df0-bdd8-4256-8183-3b340730ffa4",
"tier": "easy",
"dataset_id": "e5",
"status": "fa... | null
|
30f91c6a-f790-4e6c-afd7-cceec13bba61 | easy | viridale | 2026-08-02 11:52:01.033518+00:00 | succeeded | 7cac915827a8d442d66af67af462e8919e2ce7f653adb50c018ef612fbf38712 | 33,731 | null | """Sparse two-monomial execution plus committed open table.
hypothesis: v092 preserves e3 64/64 but loses 12.5% of its steps because the
scripted operator computes a second modular power for every active program,
while the candidate-owned beam profile shows only 172/512 are family 4.
Index-select family-4 ... | {
"id": "30f91c6a-f790-4e6c-afd7-cceec13bba61",
"created_at": "2026-08-02 11:52:01.033518+00:00",
"db_md5": "bfbc860bb87df9f9b52f2ac327a430aa",
"submitter": "priormancer",
"github_login": "viridale",
"run_id": "f81c35b9-1a79-4e5a-b658-b458e6123e5f",
"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,
"... |
31018dea-8bb7-4588-8383-49f247207cf7 | easy | khushidahi | 2026-08-20 03:31:59.202898+00:00 | succeeded | d4547591839bdc3d40724468d213e0f958b19c5375b8046a94a6abe45c3a6d5d | 31,842 | null | """R294: wide causal two-stage categorical quadratic functional root.
This candidate is deliberately diagnostic. It trains on evaluator-owned final
labels at several depths so that one shared transition must compose. The model
never converts N or X to native numbers and never implements multiplication or
modular red... | {
"id": "31018dea-8bb7-4588-8383-49f247207cf7",
"created_at": "2026-08-20 03:31:59.202898+00:00",
"db_md5": "cc085b9a3eb44171c2813cba916ec6c4",
"submitter": "khushidahi",
"github_login": "khushidahi",
"run_id": "4d8b33a9-7fde-4f34-a129-a591cd1e6afb",
"tier": "easy",
"dataset_id": "e3",
"status": "succ... | {
"score": {
"mean_loss": 2.2753540407273776,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.006250000004656613
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.253523661902736,
"example_count": 800... |
3103b76f-9e0e-4eff-9348-85cd650adf21 | easy | achyutreddy24 | 2026-08-04 23:10:56.236405+00:00 | failed | debb8997b670dd3716ccfb0f4c0547f31994d8d5f919f15b0d28afebbb3344b0 | 10,638 | null | """A looped transformer with learned, input-conditioned depth.
WHAT THIS IS
============
This submission contains no task-specific arithmetic: no modular structure, no
hand-set initialization, no token-map constants. It is a single generic
transformer block applied repeatedly to a latent state, with a learned halting... | {
"id": "3103b76f-9e0e-4eff-9348-85cd650adf21",
"created_at": "2026-08-04 23:10:56.236405+00:00",
"db_md5": "ca309437a908d354df935109d837877a",
"submitter": "Achyut Reddy",
"github_login": "achyutreddy24",
"run_id": "1dd1dc55-f00b-44eb-9539-876267347fe9",
"tier": "easy",
"dataset_id": "e1",
"status": ... | null
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.