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 |
|---|---|---|---|---|---|---|---|---|---|---|
0f760dd2-b055-434a-915d-9897bd8c1758 | easy | mnida | 2026-08-25 17:11:47.334781+00:00 | succeeded | a17f04dc6e0342a5fbd325549b57008804a0396354c83f73543bc2f5697655d7 | 15,667 | null | """Continuous-state quadratic looped Transformer with immutable routes."""
from __future__ import annotations
import math
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
assert_model_state,
)... | {
"id": "0f760dd2-b055-434a-915d-9897bd8c1758",
"created_at": "2026-08-25 17:11:47.334781+00:00",
"db_md5": "90b721b12142a7356cb668035d1ef8b7",
"submitter": "mnida",
"github_login": "mnida",
"run_id": "ab042b2f-3be4-43d7-9067-af81eba7086e",
"tier": "easy",
"dataset_id": "e5",
"status": "succeeded",
... | {
"score": {
"mean_loss": 6.491118983348401,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.004583333333333333
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 6.7500006372084,
"example_count": 600,
... |
0f762a37-2100-4aee-bc5a-b42e0b1d34d2 | easy | sapient-sapiens | 2026-08-20 20:46:46.630015+00:00 | succeeded | a02688fc26f74a1fef3e44eb71fcea9b5571b8beb79eb403da5f7614b45ee5be | 38,230 | null | """Hierarchical learned digit-mixing recurrence with an answer-only objective.
The recurrent state is a bank of right-aligned digit slots. Every operator
application is followed by a soft quantization back onto the token simplex, and
the digit logits produced by that quantization are the answer logits at that
depth. T... | {
"id": "0f762a37-2100-4aee-bc5a-b42e0b1d34d2",
"created_at": "2026-08-20 20:46:46.630015+00:00",
"db_md5": "401f470a6dfe03a40a53b9a18dc151b8",
"submitter": "Ertondy",
"github_login": "sapient-sapiens",
"run_id": "c0353d2c-db30-44c8-b03e-aa803b6683e8",
"tier": "easy",
"dataset_id": "e5",
"status": "su... | {
"score": {
"mean_loss": 2.170398634998506,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.007916666666666667
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.1737612050210413,
"example_count": 600... |
0f77e493-9f5a-4081-b36a-fceb12c2bc9f | easy | yashkant | 2026-08-17 00:24:06.223203+00:00 | succeeded | 9b4cfe8e0bf5b7cab3ca2ef38181134e6f5e70d2df37a63d1efa9605a7811cec | 43,554 | null | """Round2190 removes cross-step Dykstra correction carry.
The retained official carrier, primal recurrence, two learned projections,
antisymmetric set codes, parsed-T allocation, optimizer, and EMA are fixed.
The candidate reuses the learned initial A/B correction offsets on every
cycle instead of carrying hand-writte... | {
"id": "0f77e493-9f5a-4081-b36a-fceb12c2bc9f",
"created_at": "2026-08-17 00:24:06.223203+00:00",
"db_md5": "0754f6fe90068b218bbcff849a9e897c",
"submitter": "Yash Kant",
"github_login": "yashkant",
"run_id": "5eb14b83-dec8-4ef7-a2e4-2dfd9cfad1e9",
"tier": "easy",
"dataset_id": "e2",
"status": "succeed... | {
"score": {
"mean_loss": 2.2879951000213623,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.025000001303851604
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.3783962726593018,
"example_count": 30... |
0f84c250-20ee-4d22-b5a1-f31f62e14302 | easy | karanganesan | 2026-08-06 05:19:50.899433+00:00 | succeeded | 541970e8ce775f76824782978d1f913dccfa423608e979f5519161ce241b6909 | 25,616 | 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": "0f84c250-20ee-4d22-b5a1-f31f62e14302",
"created_at": "2026-08-06 05:19:50.899433+00:00",
"db_md5": "a871d5cc8ca91f873a0e8488dfc9a475",
"submitter": "Karan Ganesan",
"github_login": "karanganesan",
"run_id": "8f5b00b5-9f1a-4aa9-96e6-f9f0b56b577b",
"tier": "easy",
"dataset_id": "e1",
"status": ... | {
"score": {
"mean_loss": 2.3456339836120605,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.011666666716337204
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.2430825233459473,
"example_count": 10... |
0f88ae9f-7a26-4f25-8e52-6454d27cfbb6 | easy | liam-gb | 2026-08-13 01:17:50.680141+00:00 | succeeded | 53d7e91380981f1d6181d693f3d88fca7bed04b89915f1c674f62dede08ac187 | 14,798 | null | """rns_pm_gather_ct_widet_lrt30: widet with lookup tables at lr 3e-2.
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 ... | {
"id": "0f88ae9f-7a26-4f25-8e52-6454d27cfbb6",
"created_at": "2026-08-13 01:17:50.680141+00:00",
"db_md5": "7f2d2e206e1dbec836b6accd45dea1f1",
"submitter": "liam-gb",
"github_login": "liam-gb",
"run_id": "cdefc3e7-458d-4c0f-b192-c663d0d27d4f",
"tier": "easy",
"dataset_id": "e4",
"status": "succeeded"... | {
"score": {
"mean_loss": 4.64902937414953,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.025729166890184087
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 4.919563037575806,
"example_count": 1200,... |
0f8c1363-d2a0-4f7d-b7bf-f17c58f4cb46 | easy | khushidahi | 2026-08-20 02:24:49.903513+00:00 | succeeded | e896b5d9a003a39cc4a48222f217367fe6fe42fd305eb488cd5ab417a35b9f0c | 39,296 | null | """R282: short-operand bootstrap for a tied quadratic root.
This candidate targets functional-root recovery when the evaluator supplies
only composed endpoints such as T=2. It trains only from evaluator-owned final
labels and always executes the supplied recurrence depth. The model never
converts N or X to native nu... | {
"id": "0f8c1363-d2a0-4f7d-b7bf-f17c58f4cb46",
"created_at": "2026-08-20 02:24:49.903513+00:00",
"db_md5": "607e418b6edfffec20552b91e595e954",
"submitter": "khushidahi",
"github_login": "khushidahi",
"run_id": "a4066991-d9a2-48a9-9c62-d58835bffd1b",
"tier": "easy",
"dataset_id": "e3",
"status": "succ... | {
"score": {
"mean_loss": 2.212706284862254,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.004375
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.202818362713863,
"example_count": 800,
"... |
0f938c02-f8d4-4562-b2f8-336757659239 | easy | yashkant | 2026-08-14 16:10:02.598638+00:00 | succeeded | d446d91ed5abce6dac8ae986d9f8493f6ac1c70035a82a0aff6cdf4306e3b6e8 | 51,823 | null | """Round1961 field-local positioned-pair hash conditioning.
The complete Round1939 T-field-0.025 exact-600 carrier remains intact. One
zero-start learned table is addressed by categorical adjacent digit pairs,
their right-relative local position class, and the public N-versus-X field
tag. Separate N/X means are conc... | {
"id": "0f938c02-f8d4-4562-b2f8-336757659239",
"created_at": "2026-08-14 16:10:02.598638+00:00",
"db_md5": "78058de8ca076c3e7fc71fadd5aa2f09",
"submitter": "Yash Kant",
"github_login": "yashkant",
"run_id": "4aa64543-75a9-4f0a-aae8-44298b318b9d",
"tier": "easy",
"dataset_id": "e1",
"status": "succeed... | {
"score": {
"mean_loss": 5.711158275604248,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.08166666701436043
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 6.850822925567627,
"example_count": 100,
... |
0f948f27-acf9-4b4c-97c8-daff980405d5 | easy | blake-camp-surge | 2026-08-05 19:14:28.321736+00:00 | succeeded | 308e832c8cee7cb1aac2bd298c2e16f0dfaf840aca4d2dede08fe826427dbeff | 10,518 | null | """Vector-state neural cellular automaton over attention edges."""
from __future__ import annotations
import math
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
assert_model_state,
)
TOKEN... | {
"id": "0f948f27-acf9-4b4c-97c8-daff980405d5",
"created_at": "2026-08-05 19:14:28.321736+00:00",
"db_md5": "72f2e4ded4a7c3fe6f89f854ec1820f2",
"submitter": "Blake Camp",
"github_login": "blake-camp-surge",
"run_id": "cf86600c-9685-492b-b861-c6f709ff970b",
"tier": "easy",
"dataset_id": "e3",
"status":... | {
"score": {
"mean_loss": 2.1290325061776674,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.00375
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.1274903352189374,
"example_count": 800,
... |
0f963796-e461-45b4-8999-e5c218a104e2 | easy | richardcepka | 2026-08-13 20:17:46.504494+00:00 | succeeded | 28d9b030305124619790cb81aa390015ddbed19116732d66249e955e7a5b559e | 23,599 | null | """Recurrent digit-register model for One Layer Deeper.
A looped transformer whose recurrent state is a decimal digit register that is
re-quantised through a ten-entry codebook on every iteration. Nothing inside
the loop depends on the iteration index and no parameter is indexed by an
absolute slot position, so the e... | {
"id": "0f963796-e461-45b4-8999-e5c218a104e2",
"created_at": "2026-08-13 20:17:46.504494+00:00",
"db_md5": "5d36ff4f3a7f3de2e74c519dcb5ddc5f",
"submitter": "Richard Cepka",
"github_login": "richardcepka",
"run_id": "974f5bff-04e1-4791-bcfc-00fbe4138eaf",
"tier": "easy",
"dataset_id": "e1",
"status": ... | {
"score": {
"mean_loss": 2.2094438076019287,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0833333320915699
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.2315218448638916,
"example_count": 100,... |
0f984769-479a-42f9-8554-68d49f9c9ec9 | easy | yashkant | 2026-08-12 04:32:10.900877+00:00 | succeeded | 72589ccbf906c8b4391b37448566aec91cb84915dd3e52536a2ddceaed8d62e4 | 31,390 | null | """Round1702 one-projection-per-T Dykstra recursion on the D416 carrier.
Each parsed public-T cycle applies one shared learned smooth projection and
alternates the A/B set code and correction memory across cycles. The immutable
anchor is reinjected on every call; all active cycle outputs are averaged, and
terminal CE... | {
"id": "0f984769-479a-42f9-8554-68d49f9c9ec9",
"created_at": "2026-08-12 04:32:10.900877+00:00",
"db_md5": "c6650101fcbfb57fcf623704835e23f7",
"submitter": "Yash Kant",
"github_login": "yashkant",
"run_id": "e21e4062-dcc6-415e-936b-abfafc0d9cfd",
"tier": "easy",
"dataset_id": "e2",
"status": "succeed... | {
"score": {
"mean_loss": 5.059478521347046,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.3812500275671482
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 8.711710929870605,
"example_count": 300,
... |
0fa0feed-5137-4d24-ae0f-b654ea3de93a | easy | jordanrubin | 2026-08-10 20:26:21.323741+00:00 | succeeded | d66d3596d565e364e36b2e65e3e2d2e8d5bb226368b36c4f9dd3d714e86dfa54 | 19,332 | 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": "0fa0feed-5137-4d24-ae0f-b654ea3de93a",
"created_at": "2026-08-10 20:26:21.323741+00:00",
"db_md5": "881e1109fb87d3faf946c8d57e92a5d3",
"submitter": "Jordan Rubin",
"github_login": "jordanrubin",
"run_id": "f3df5963-711d-4956-a7a2-2ac1f3a7aefd",
"tier": "easy",
"dataset_id": "e5",
"status": "s... | {
"score": {
"mean_loss": 3.043394365884261,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.00458333323088785
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.197231356338535,
"example_count": 600,
... |
0fa86066-e886-434b-b956-c8b3e87d9238 | easy | khushidahi | 2026-08-05 22:06:50.092370+00:00 | succeeded | b9856becae4169712872610cba7120737852309a157f8dd5b91b8bd17983ea99 | 25,646 | null | """R16 hard-transfer hybrid experiments for One Layer Deeper.
Motivated by complementary Hard behavior:
- the generic sequence recurrent model retained seen-N first-rung progress;
- the structured decimal-tape model retained OOD-N first-rung progress.
The dual variants co-train both learned representations, supervise... | {
"id": "0fa86066-e886-434b-b956-c8b3e87d9238",
"created_at": "2026-08-05 22:06:50.092370+00:00",
"db_md5": "bab36deb022bdeac259a5ed721cf495b",
"submitter": "khushidahi",
"github_login": "khushidahi",
"run_id": "a6d5dab3-9672-4816-a95c-e6e2073a350c",
"tier": "easy",
"dataset_id": "e5",
"status": "succ... | {
"score": {
"mean_loss": 2.190055696075564,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.011666666896392902
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.192997694015503,
"example_count": 600,... |
0fb9aeae-17a3-48cb-b56a-74c8caa8059d | easy | yashkant | 2026-08-17 01:08:34.719420+00:00 | succeeded | 40184f8861c4649808c05e9ddd3bfafbca29c57f35aff111085ac5da1984191b | 49,530 | null | """Round2217 adds a zero-start square/reduce/answer recurrence.
The confirmed correction-reset model remains intact. Its width-8
answer state is refined by one shared transition with explicit product,
reduction, and answer roles. Immutable categorical N and X summaries
are reinjected on every call and physical calls... | {
"id": "0fb9aeae-17a3-48cb-b56a-74c8caa8059d",
"created_at": "2026-08-17 01:08:34.719420+00:00",
"db_md5": "f17b945fa5b3863a5efdac21e4eba689",
"submitter": "Yash Kant",
"github_login": "yashkant",
"run_id": "576d255e-d98c-4072-ac20-4e0d475e31c9",
"tier": "easy",
"dataset_id": "e5",
"status": "succeed... | {
"score": {
"mean_loss": 2.1497453210262636,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.007500000012417634
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.1542129185167664,
"example_count": 60... |
0fc67334-e3b5-44b0-87c2-d9fffcb77427 | easy | sapient-sapiens | 2026-08-04 17:19:49.557978+00:00 | succeeded | 1c213700c8e6b75e4a74895cb319528b8b7842928d8c36cda059349abf08c13f | 18,639 | null | """v2: weight-tied recurrent Transformer with depth-aware supervision.
Fixes relative to Phase 2–6:
- Per-row T parse; supervise only rows with T <= depth at each readout.
- Genuine hold terms beyond T; deep-curriculum under truncated BPTT.
- Power-iteration spectral normalization (contractive residual).
- Det... | {
"id": "0fc67334-e3b5-44b0-87c2-d9fffcb77427",
"created_at": "2026-08-04 17:19:49.557978+00:00",
"db_md5": "774262f02d5bb32589cb611db1a14ce4",
"submitter": "Ertondy",
"github_login": "sapient-sapiens",
"run_id": "b3a77331-961e-421e-9c4f-1d9d96c065bf",
"tier": "easy",
"dataset_id": "e5",
"status": "su... | {
"score": {
"mean_loss": 2.3405789543065594,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.005
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.3624077846918405,
"example_count": 600,
"e... |
0fc6c381-ef70-4c17-af91-8416869fe1b9 | easy | jordanrubin | 2026-08-02 18:24:53.945182+00:00 | succeeded | f8981ce5397863750ef37a1938203c582fac1d301714ad11fabf35606f0f6569 | 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": "0fc6c381-ef70-4c17-af91-8416869fe1b9",
"created_at": "2026-08-02 18:24:53.945182+00:00",
"db_md5": "760eb467b92c185a8c3c640efafc1500",
"submitter": "Jordan Rubin",
"github_login": "jordanrubin",
"run_id": "b385c56f-4e9e-4178-a39e-a62f25c23212",
"tier": "easy",
"dataset_id": "e1",
"status": "s... | {
"score": {
"mean_loss": 4.46888713666821,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.18333332675198713
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 6.75981898628064,
"example_count": 100,
... |
0fc81aca-62ff-4a6d-a1cc-cd7db72ba68a | easy | shan23chen | 2026-08-08 02:36:33.606401+00:00 | succeeded | fecc57ffa78422b4a21783c166a8670ae97878bc1c0db4c1a330b875c043909e | 20,431 | null | """Boundary-reset recurrent register Transformer.
One learned transition is applied once per requested outer step. The
transition receives an immutable decimal N register and the current learned
boundary state, but never receives T or a separate copy of the original X.
Within each outer application, scratch persists ... | {
"id": "0fc81aca-62ff-4a6d-a1cc-cd7db72ba68a",
"created_at": "2026-08-08 02:36:33.606401+00:00",
"db_md5": "1f0d9ea3966c06e41a72df6ba36dc0c6",
"submitter": "Shan Chen",
"github_login": "shan23chen",
"run_id": "a2b23f7a-b4cf-48ad-8534-c171f5faaed7",
"tier": "easy",
"dataset_id": "e5",
"status": "succe... | {
"score": {
"mean_loss": 2.496020793914795,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.008333333535119891
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.3050479888916016,
"example_count": 600... |
0fca12d4-0f57-4d32-bbe1-b9bb07c0d7bc | easy | jordanrubin | 2026-08-07 06:08:00.384490+00:00 | succeeded | 6f17eb27b2bb1fbb5891f0392535694f326c447b00cd9f37136d01d78d2e30ec | 16,908 | null | """Autonomous digit-state recurrent Transformer for repeated modular squaring.
The model is deliberately organized around one learned transition:
p_0 = right_aligned_decimal_digits(x)
p_{k+1} = F_theta(p_k, decimal_digits(N))
The same two-block Transformer cell is applied exactly T times. T is used only
as ... | {
"id": "0fca12d4-0f57-4d32-bbe1-b9bb07c0d7bc",
"created_at": "2026-08-07 06:08:00.384490+00:00",
"db_md5": "378594dd4459a9c7efa73ec8d9805eae",
"submitter": "Jordan Rubin",
"github_login": "jordanrubin",
"run_id": "acb7d67b-bfec-4cf2-9efb-75a264fd848d",
"tier": "easy",
"dataset_id": "e5",
"status": "s... | {
"score": {
"mean_loss": 4.466879699483501,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.007916666489715379
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 4.551595894878755,
"example_count": 600,... |
0fd8c0f9-3030-4b22-a41c-4f47c715c802 | easy | Dandandan | 2026-08-05 17:01:40.703321+00:00 | failed | 48fb1946e463a1a5e84e1e265ce19e72b8d28c6794268e7699b549ceb5ece828 | 33,995 | null | """End-to-end recursive function learner for One Layer Deeper.
A population of randomly initialized polynomial register networks is reused
exactly once per requested outer step. The evaluator trains every coefficient
and the evidence selector through its normal forward/backward/optimizer loop.
There is no recurrence-... | {
"id": "0fd8c0f9-3030-4b22-a41c-4f47c715c802",
"created_at": "2026-08-05 17:01:40.703321+00:00",
"db_md5": "eea772f609f6ea00d9b9530252d54942",
"submitter": "Daniël Heres",
"github_login": "Dandandan",
"run_id": "bfeac944-ff12-415a-9ad5-2b7b57ffff03",
"tier": "easy",
"dataset_id": "e5",
"status": "fai... | null
|
0fd95ce8-6755-449a-bdff-c2cfd0bd16bc | easy | erdavis0 | 2026-08-19 08:29:11.388891+00:00 | succeeded | 3e90438ca6f93e0fa30aa159e6c8ddf622ebbc8c0c6803d879e9d68c8690e9d2 | 10,891 | null | """Field-separable raw-vocabulary recurrent operator with identity decoding."""
from __future__ import annotations
import math
import time
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
Toke... | {
"id": "0fd95ce8-6755-449a-bdff-c2cfd0bd16bc",
"created_at": "2026-08-19 08:29:11.388891+00:00",
"db_md5": "35ac0dd3de7284e050e495eb8ee0c982",
"submitter": "Ethan",
"github_login": "erdavis0",
"run_id": "76023f8d-0a33-492e-b440-39a640d56720",
"tier": "easy",
"dataset_id": "e3",
"status": "succeeded",... | {
"score": {
"mean_loss": 2.271347065251822,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.018125000000000002
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.2708335813474996,
"example_count": 800... |
0fe255c2-b418-4f4f-93a4-88449d833334 | easy | oupadhyay | 2026-08-29 19:00:29.746399+00:00 | succeeded | 4eab258361c06b3ff627450c285de21fb0641babb68b8eedb474faf260a36059 | 4,276 | null | """Conditional structured-loss phase 5: focal125."""
from __future__ import annotations
import torch
import torch.nn.functional as F
from torch import nn
from benchmark import ModelSpec,OptimizerSpec,OptimizerBundle,Submission,TokenLossBatch,assert_model_state
W,D,H=4,64,32
class C:
def __init__(s,vocab_size,m... | {
"id": "0fe255c2-b418-4f4f-93a4-88449d833334",
"created_at": "2026-08-29 19:00:29.746399+00:00",
"db_md5": "7bc218d28535f75b7ee24d95dbdabff6",
"submitter": "Ojasw Upadhyay",
"github_login": "oupadhyay",
"run_id": "997b0654-f52d-4058-af91-90d6682ac7ae",
"tier": "easy",
"dataset_id": "e6",
"status": "s... | {
"score": {
"mean_loss": 1.9638167023658752,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.05022522807121277
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 1.928548812866211,
"example_count": 60,
... |
0fe83495-4d60-4aca-813c-bed141f4d3e6 | easy | mcleish7 | 2026-08-15 01:50:44.339331+00:00 | succeeded | 7bf30e19d90e7f6cadb4ad67d0b2cd51634a4fd2d29a507be6f840a0dd78f8c1 | 3,628 | null | """Basic single-pass Transformer with PyTorch AdamW."""
from __future__ import annotations
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
assert_model_state,
)
D_MODEL = 128
NUM_HEADS = 4
... | {
"id": "0fe83495-4d60-4aca-813c-bed141f4d3e6",
"created_at": "2026-08-15 01:50:44.339331+00:00",
"db_md5": "eaac69fe345f82973fcf392c504a40ff",
"submitter": "Sean McLeish",
"github_login": "mcleish7",
"run_id": "9545fa46-134a-4c4a-9266-4164693f7659",
"tier": "easy",
"dataset_id": "e1",
"status": "succ... | {
"score": {
"mean_loss": 3.036493420600891,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.011666666716337204
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 3.965740203857422,
"example_count": 100,... |
0fffd29f-5974-4851-9f51-5da2eb4c55a4 | easy | ipsa-loquitur | 2026-08-26 04:49:24.597481+00:00 | succeeded | e0ae04a6d1a7513b7c75f7b7851245b42873696e46567183197b58178157f8ae | 15,930 | null | """Standalone NetF-V1 submission candidate for One Layer Deeper.
The learned rule is shared across bit positions, modular-square iterations,
moduli, and tape widths. During training, exact double-and-add states provide
dense one-step supervision. During evaluation, the model receives no state
corrections and rolls out... | {
"id": "0fffd29f-5974-4851-9f51-5da2eb4c55a4",
"created_at": "2026-08-26 04:49:24.597481+00:00",
"db_md5": "23b55787c8a66f7af6e9f7a59ff8a21d",
"submitter": "ipsa-loquitur",
"github_login": "ipsa-loquitur",
"run_id": "e230358b-8cd5-467b-9c76-d08218624724",
"tier": "easy",
"dataset_id": "e4",
"status":... | {
"score": {
"mean_loss": 16.187870488500838,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.2195833338300387
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 18.5590435682866,
"example_count": 1200,
... |
1007a94d-94b7-4bf7-8a3d-fbd27db77732 | easy | oupadhyay | 2026-08-28 02:11:00.922290+00:00 | succeeded | 5aa53a7497ec7e868fc0a42bbef086dfabd32bfe4a431001e91f2044a96e4ee1 | 4,297 | null | """Dynamic D64 conditional transducer with pure T1 token weighting."""
from __future__ import annotations
import torch
import torch.nn.functional as F
from torch import nn
from benchmark import ModelSpec,OptimizerSpec,OptimizerBundle,Submission,TokenLossBatch,assert_model_state
W,D,H=None,64,32
class C:
def __init__(s... | {
"id": "1007a94d-94b7-4bf7-8a3d-fbd27db77732",
"created_at": "2026-08-28 02:11:00.922290+00:00",
"db_md5": "cdc0dd364399e103fe0294bfda5cea10",
"submitter": "Ojasw Upadhyay",
"github_login": "oupadhyay",
"run_id": "d7446677-dc28-4c15-a2d1-8f012ee52330",
"tier": "easy",
"dataset_id": "e6",
"status": "s... | {
"score": {
"mean_loss": 2.002084255218506,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.09391892328858376
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 1.9658660888671875,
"example_count": 60,
... |
1018bf2a-87c4-4d88-a3ff-b4f3ea1ab090 | easy | erdavis0 | 2026-08-19 07:43:32.135348+00:00 | succeeded | d3d07b0f1e5e1cf681cb72abdc236ee3f077e418f32e40c9a73b761ed0abbee7 | 7,281 | null | """Generic bidirectional encoder with a recurrent soft-token output decoder."""
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,
asse... | {
"id": "1018bf2a-87c4-4d88-a3ff-b4f3ea1ab090",
"created_at": "2026-08-19 07:43:32.135348+00:00",
"db_md5": "21ef16843e31305096cff1a7c0eba857",
"submitter": "Ethan",
"github_login": "erdavis0",
"run_id": "c706935a-2619-43f5-8561-714db52f529c",
"tier": "easy",
"dataset_id": "e1",
"status": "succeeded",... | {
"score": {
"mean_loss": 5.457361459732056,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.04166666604578495
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 5.522273063659668,
"example_count": 100,
... |
1018ee11-7c26-497c-97d3-2a1403b81d28 | easy | repst | 2026-08-22 22:03:27.101605+00:00 | succeeded | 16dd77f86d5b850daaaff2dca807582982d515e1c3091041bd18ba9f97879062 | 16,702 | null | # R524: latent expression-program recurrent machine.
#
# A learned autoregressive policy selects an exact expression tree and anonymous
# field-role binding from ordinary output-token loss. The authored interpreter
# applies that stationary expression for T steps and a learned digit lexicon
# serializes the result. U... | {
"id": "1018ee11-7c26-497c-97d3-2a1403b81d28",
"created_at": "2026-08-22 22:03:27.101605+00:00",
"db_md5": "6ceb9720a359976ec6b1ad81564fd561",
"submitter": "Ryan Epstein",
"github_login": "repst",
"run_id": "b6c4973b-4dc8-41f6-a93e-ae95829f60bc",
"tier": "easy",
"dataset_id": "e1",
"status": "succeed... | {
"score": {
"mean_loss": 0.00012529729428933933,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 1.0
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 0.00012255726323928684,
"example_count": 100,
... |
102098f4-8fa6-454e-8154-90d416e79c4a | easy | ehonig | 2026-08-14 01:21:12.603417+00:00 | succeeded | d4863f4de678d842869b640fbf6ce59ff6dec0b2ff6dff8337c93025539210f1 | 54,753 | null | """DigitLoop — a place-value recurrent model for repeated modular squaring.
The task is ``y = x^(2^T) mod N`` with the answer read back tail aligned, one
decimal place value per token position from the right. A stack of attention and
MLP layers has no primitive for the two operations this needs — a partial-product
co... | {
"id": "102098f4-8fa6-454e-8154-90d416e79c4a",
"created_at": "2026-08-14 01:21:12.603417+00:00",
"db_md5": "f62c4fc3d75fd4327b86be1aebf443ce",
"submitter": "Edouardo Honig",
"github_login": "ehonig",
"run_id": "a344f55b-e65e-484e-8f46-979ca4f0eb32",
"tier": "easy",
"dataset_id": "e1",
"status": "succ... | {
"score": {
"mean_loss": 5.610818386077881,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.19499999657273293
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 7.214132308959961,
"example_count": 100,
... |
1020a7f9-b1ab-416e-a9dd-d5a61a318cc8 | easy | DDanlov | 2026-08-15 21:12:58.648016+00:00 | succeeded | 3eb517cb9667863720ce75d439fd67ebc36b79a26534e0b145dad12284313550 | 17,950 | null | """
Official Submission for One Layer Deeper Challenge
Architecture: Deep Equilibrium Port-Hamiltonian Deliberation Network (Model C v3 / DEPH-Net)
- Input: Multi-Scale Linear Harmonic Torus Frequency Encoder (omega_j = 1..64)
- Recurrent Core: Continuous Port-Hamiltonian Dynamical Flow dz/dt = (J - R) * grad_E
* J =... | {
"id": "1020a7f9-b1ab-416e-a9dd-d5a61a318cc8",
"created_at": "2026-08-15 21:12:58.648016+00:00",
"db_md5": "623970f9d216ffb45b2ef4f8507de941",
"submitter": "DDanlov",
"github_login": "DDanlov",
"run_id": "2af8ce55-0bb8-41d4-ba95-8dac8efcb82d",
"tier": "easy",
"dataset_id": "e5",
"status": "succeeded"... | {
"score": {
"mean_loss": 2.412507358349103,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.009166666679084301
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.417765759833725,
"example_count": 600,... |
1020b452-ac21-40fa-8136-130fda1d2349 | easy | velocizapkar | 2026-08-16 07:13:23.322758+00:00 | succeeded | f9b8cc409b413dfc820863038872db237130cacbed1b67e58c60d1474c445e83 | 15,294 | null | """Tied skip-free Transformer with a wall-time depth curriculum.
Training begins with one application of the two-pair shared group and advances
to four repeats using only evaluator-provided wall time. Evaluation always
uses the donor's full four repeats. Parameters, initialization, topology,
AdamW, token loss, batch... | {
"id": "1020b452-ac21-40fa-8136-130fda1d2349",
"created_at": "2026-08-16 07:13:23.322758+00:00",
"db_md5": "20868af55f4392a1b7b265fd3c8a48ff",
"submitter": "Aakanksh Zarapkar",
"github_login": "velocizapkar",
"run_id": "72713e6b-4d6c-458e-b437-65b675a43a02",
"tier": "easy",
"dataset_id": "e6",
"statu... | {
"score": {
"mean_loss": 7.311941623687744,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.04842342482879758
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 6.598480224609375,
"example_count": 60,
... |
1028caa4-88e1-4a89-af62-b3dd9b49b131 | easy | yashkant | 2026-08-12 04:34:31.698148+00:00 | succeeded | 46a072dfc72775006aadcbca85d1a443c707401c656e9b257202c9ab675b6407 | 34,480 | null | """Round1702 one-projection-per-T Dykstra recursion on the D416 carrier.
Each parsed public-T cycle applies one shared learned smooth projection and
alternates the A/B set code and correction memory across cycles. The immutable
anchor is reinjected on every call; all active cycle outputs are averaged, and
terminal CE... | {
"id": "1028caa4-88e1-4a89-af62-b3dd9b49b131",
"created_at": "2026-08-12 04:34:31.698148+00:00",
"db_md5": "14be733ac22a5dc546702dbe4fe15d84",
"submitter": "Yash Kant",
"github_login": "yashkant",
"run_id": "587a1447-79d9-4cfd-9cb6-e89f6568d9ab",
"tier": "easy",
"dataset_id": "e3",
"status": "succeed... | {
"score": {
"mean_loss": 2.123806821861843,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.016250000121071934
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.123247481214935,
"example_count": 800,... |
102d60c6-3cf3-44cc-8e40-d107eb90e00a | easy | yashkant | 2026-08-16 00:41:02.403957+00:00 | succeeded | 106c18dc75358f3b836b9a461b2a4eb0ccc23c93e01ce91f9872a20fe8ad8590 | 38,959 | null | """Round1818 parent-preserving temporal Anderson-Dykstra mixing.
The complete Round1792 AdamW5/EMA.995 W8/D2 carrier remains byte-identical at
a matched seed. Initialization of the added temporal mixer is isolated from
the inherited CPU RNG stream, and one exact-zero scalar gate keeps the scored
path equal to the par... | {
"id": "102d60c6-3cf3-44cc-8e40-d107eb90e00a",
"created_at": "2026-08-16 00:41:02.403957+00:00",
"db_md5": "4afb933fc29047bb02333e1b1246a013",
"submitter": "Yash Kant",
"github_login": "yashkant",
"run_id": "1b7d9a78-752d-4fb1-a3a4-50369e9473b8",
"tier": "easy",
"dataset_id": "e5",
"status": "succeed... | {
"score": {
"mean_loss": 2.145938853480141,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0079166666790843
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.1510462953370784,
"example_count": 600,
... |
10307ef0-e32c-4d17-ad5c-7992690f397c | easy | liam-gb | 2026-08-17 11:43:13.088873+00:00 | succeeded | 936ea05de67f15515fc3e81d838b067bcbeb44d73d467fc5f6b2156fc87ef97a | 15,810 | null | """rulemix: the per-prime step is a LEARNED MIXTURE OVER FIXED EXPONENT MAPS.
The identification-transfer body. Its predecessor lineage has per-class
freedom in the transition (pointwise pushforwards can be fit row by row), so
nothing forces uncovered residue classes to obey the rule identified on
covered ones. Here t... | {
"id": "10307ef0-e32c-4d17-ad5c-7992690f397c",
"created_at": "2026-08-17 11:43:13.088873+00:00",
"db_md5": "be7d4a810b4a0f98b680749deda31df0",
"submitter": "liam-gb",
"github_login": "liam-gb",
"run_id": "19e9986d-a23a-4ec0-ae4b-c09673ffefe3",
"tier": "easy",
"dataset_id": "e2",
"status": "succeeded"... | {
"score": {
"mean_loss": 2.3745601177215576,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.7100000381469727
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.624119758605957,
"example_count": 300,
... |
1033ae25-2ba9-45da-990a-4bc7cbcf965d | easy | DDanlov | 2026-08-18 08:57:22.522128+00:00 | succeeded | 2daa3dabbf47e1f004651c2e7895944312e775b363dd9a9f753435d4e887457f | 23,511 | null | """
Official Submission for One Layer Deeper Challenge
Architecture: 4-Layer Recurrent Core Pure No-Norm No-Skip DKS Reasoning Model:
- 4 Distinct Pure No-Norm No-Skip Transformer Layers in the Recurrent Core: L1 -> L2 -> L3 -> L4
- Unrolled T times for serial modular squaring
- dim=512, 16 heads, d_ff=1024, RohanShamp... | {
"id": "1033ae25-2ba9-45da-990a-4bc7cbcf965d",
"created_at": "2026-08-18 08:57:22.522128+00:00",
"db_md5": "a45a27d6e56da3f75f753d1345b5cf43",
"submitter": "DDanlov",
"github_login": "DDanlov",
"run_id": "dae0e255-a9b1-4003-b55a-28eb780badf0",
"tier": "easy",
"dataset_id": "e1",
"status": "succeeded"... | {
"score": {
"mean_loss": 2.480574579802748,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.05500000149011612
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.424070440121551,
"example_count": 100,
... |
1035bbd8-1dd4-4708-8db9-234a773ad180 | easy | oupadhyay | 2026-08-29 18:56:44.469384+00:00 | succeeded | 6aabe7f902eafd72bca8518a0bdf10a6b3c7ebf090db2afed5eb920fed202c16 | 4,273 | null | """Conditional structured-loss screen: focal20."""
from __future__ import annotations
import torch
import torch.nn.functional as F
from torch import nn
from benchmark import ModelSpec,OptimizerSpec,OptimizerBundle,Submission,TokenLossBatch,assert_model_state
W,D,H=4,64,32
class C:
def __init__(s,vocab_size,max... | {
"id": "1035bbd8-1dd4-4708-8db9-234a773ad180",
"created_at": "2026-08-29 18:56:44.469384+00:00",
"db_md5": "255b8b3d7089ab4181b16c174feeaf6a",
"submitter": "Ojasw Upadhyay",
"github_login": "oupadhyay",
"run_id": "7735e6b4-1587-4114-8856-32ea65c8e9bb",
"tier": "easy",
"dataset_id": "e6",
"status": "s... | {
"score": {
"mean_loss": 1.9988753199577332,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.052027029916644096
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 1.9893947839736938,
"example_count": 60... |
1039598c-987a-4541-874a-6934d90cad11 | easy | lpbb | 2026-08-18 10:17:52.817944+00:00 | succeeded | e2a1d27b63561dc4daf76a27b9f4711115489649b9e039daca976ab84a4d02d1 | 6,792 | null | """S0 + abacus embeddings: same 12 untied layers as s0, plus a per-digit
significance tag (units=1, tens=2, ...) embedded additively, with random
per-row offsets at train time so tag rows beyond the trained digit counts
still receive gradient (McLeish et al. 2024)."""
from __future__ import annotations
import torch
i... | {
"id": "1039598c-987a-4541-874a-6934d90cad11",
"created_at": "2026-08-18 10:17:52.817944+00:00",
"db_md5": "779b65fa0f4c1ab25f4d6053add506de",
"submitter": "Lpbb",
"github_login": "lpbb",
"run_id": "ceba3544-1d48-434a-9e4a-856c28bf9f64",
"tier": "easy",
"dataset_id": "e1",
"status": "succeeded",
"s... | {
"score": {
"mean_loss": 6.183819055557251,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.024999999441206455
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 6.074984550476074,
"example_count": 100,... |
103b7337-a9de-4f76-88b9-2f49a7826793 | easy | karanganesan | 2026-08-05 11:34:23.398637+00:00 | succeeded | 02093a1147e7e0f7a6f31766372ba87b210a90ed89ba01c8f61e95b41e1ed1db | 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": "103b7337-a9de-4f76-88b9-2f49a7826793",
"created_at": "2026-08-05 11:34:23.398637+00:00",
"db_md5": "c50f3743607543d13ce21324ec5b1f2b",
"submitter": "Karan Ganesan",
"github_login": "karanganesan",
"run_id": "c03313a5-f783-4445-bd70-776c98dd006f",
"tier": "easy",
"dataset_id": "e1",
"status": ... | {
"score": {
"mean_loss": 14.263788223266602,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.044999999925494194
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 17.69736099243164,
"example_count": 100... |
1048e909-1d28-4f40-9641-a9d69c06245d | easy | shirvani-jr | 2026-08-22 13:28:14.929618+00:00 | succeeded | 0c6dbdc495254e0bce7220529f7d2fb3cede38c5c2805a4e8d7a5defe87a5b40 | 24,165 | null | """Looped exact-arithmetic digit machine.
The recurrent state is the decimal digit string of the current residue. One
loop application performs one modular squaring, so depth extrapolates in T by
construction. Every functional primitive — digit-by-digit product, full
adder, full subtractor, digit comparator, and com... | {
"id": "1048e909-1d28-4f40-9641-a9d69c06245d",
"created_at": "2026-08-22 13:28:14.929618+00:00",
"db_md5": "484dcf2452d8db1e8c365a23fff90dcf",
"submitter": "Ali",
"github_login": "shirvani-jr",
"run_id": "fa656616-be92-4e43-a377-eae9cfa9e806",
"tier": "easy",
"dataset_id": "e7",
"status": "succeeded"... | {
"score": {
"mean_loss": 0.025510301813483238,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 1.0
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 0.02552816830575466,
"example_count": 85,
"e... |
104a1e45-1939-4a84-96eb-7c5816312365 | easy | viridale | 2026-08-02 12:13:02.513571+00:00 | succeeded | fcd40f5224ef87dc006a61ede621b6df791155f0b232b04c6fa9b88f948cc6ac | 33,687 | null | """Absorbing committed table plus sparse two-monomial execution.
hypothesis: v093's persistent table_committed latch is not action-absorbing:
forward still follows the live route-logit argmax after commitment, so a
later reverse evidence run can silently restore the closed catalogue.
Once the cumulative ma... | {
"id": "104a1e45-1939-4a84-96eb-7c5816312365",
"created_at": "2026-08-02 12:13:02.513571+00:00",
"db_md5": "1a8cfeb3489262d2da9f913ccc8572ec",
"submitter": "priormancer",
"github_login": "viridale",
"run_id": "c3ae119f-ddee-40af-bf83-7577a2fbd2e6",
"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,
"... |
104a20d4-85f0-4439-afcd-57f310778609 | easy | viridale | 2026-08-28 10:42:11.122582+00:00 | succeeded | 0d2e8abe40d97d8c204c1a06f455522e66cd49a2b65eccb71ac531de47f0556e | 32,091 | null | """Implicit-lattice group-deduplicated exact-evidence slab.
hypothesis: v972's 191M-scalar candidate table is exactly the Cartesian product of
one signed 18-bit coordinate and two signed nine-value coordinates in each of three
families. It contains no learned information and is repeatedly read only to recover
... | {
"id": "104a20d4-85f0-4439-afcd-57f310778609",
"created_at": "2026-08-28 10:42:11.122582+00:00",
"db_md5": "669cdd91091184de92a5f00d38fdc844",
"submitter": "priormancer",
"github_login": "viridale",
"run_id": "ee543286-385d-4926-94ad-a900936811a4",
"tier": "easy",
"dataset_id": "e1",
"status": "succe... | {
"score": {
"mean_loss": 0.0,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 1.0
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 0.0,
"example_count": 100,
"exact_accuracy": 1.0,
"... |
104b00fe-29f9-42d3-8a12-27d130260e8f | easy | khushidahi | 2026-08-12 21:16:17.130408+00:00 | succeeded | ee0138bfdf7daa03e753345fb04f2652207071955bb8def0ef54acc4138f1591 | 16,710 | null | """R6 structured decimal-workspace model for One Layer Deeper.
The model does not implement multiplication, modular reduction, or the public
recurrence. It uses the public tokenizer structure to build learned, right-
aligned decimal tapes for N, X, and T, then applies a shared neural transition
to a mutable work tape.... | {
"id": "104b00fe-29f9-42d3-8a12-27d130260e8f",
"created_at": "2026-08-12 21:16:17.130408+00:00",
"db_md5": "1e0f597c07f0798464d5f2700531df31",
"submitter": "khushidahi",
"github_login": "khushidahi",
"run_id": "bf7e0f97-c274-4dbc-8f60-c8cf66101008",
"tier": "easy",
"dataset_id": "e5",
"status": "succ... | {
"score": {
"mean_loss": 2.222488760948181,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.012083333916962147
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.221315860748291,
"example_count": 600,... |
104b7c63-5434-4e0c-a192-16e73521b8fb | easy | jordanrubin | 2026-08-11 00:36:29.005148+00:00 | succeeded | 6311b17aab2dc50f62774ab5dd57da7d1afd4904b0baca7f45082f752f4aef82 | 20,108 | 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": "104b7c63-5434-4e0c-a192-16e73521b8fb",
"created_at": "2026-08-11 00:36:29.005148+00:00",
"db_md5": "a9dc11057225060496eae5b39f62b9fc",
"submitter": "Jordan Rubin",
"github_login": "jordanrubin",
"run_id": "9b60420f-dda6-407c-a928-5a8ef90a3bb2",
"tier": "easy",
"dataset_id": "e5",
"status": "s... | {
"score": {
"mean_loss": 2.781753955072341,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.007916666489715377
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.1903220195941326,
"example_count": 600... |
10517927-c039-4aee-9a5c-59c1493b682a | easy | DDanlov | 2026-08-17 10:36:50.222389+00:00 | succeeded | bbdd79a484dff0eb25bb8251be845dce425258b607c9b199e0ebd3fe6bfdfde7 | 28,901 | null | """
Scalable Recurrent Deliberation Transformer (ScalableDEPHNet)
============================================================
Architecture:
- Pure No-Norm: Single unit RMS normalization strictly isolated to the input token embedding layer.
Zero LayerNorm / RMSNorm inside recurrent deliberation blocks.
- Pure No-Skip... | {
"id": "10517927-c039-4aee-9a5c-59c1493b682a",
"created_at": "2026-08-17 10:36:50.222389+00:00",
"db_md5": "934ddc68dd12922853bad9512ca0a5aa",
"submitter": "DDanlov",
"github_login": "DDanlov",
"run_id": "e3a9525b-54d7-48ed-9583-ffbe1917d2fa",
"tier": "easy",
"dataset_id": "e1",
"status": "succeeded"... | {
"score": {
"mean_loss": 1.9793377062329034,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.013333333432674408
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.0614967346191406,
"example_count": 10... |
1052f619-31bc-4990-9d08-eb323db40e3b | easy | rteehas | 2026-08-19 10:08:41.189132+00:00 | succeeded | 1193e3f7fd5da8031255af24bdf6f693975d406ef4ace02bd8009ca8046acea9 | 50,839 | null | """Anonymous marginal-program submission.
Mechanism
---------
The model is a differentiable probabilistic program. Its forward pass
computes, exactly, the answer-token predictions of a small structured latent
hypothesis space:
(prompt segmentation = which token types head the three fields)
x (field-role permutat... | {
"id": "1052f619-31bc-4990-9d08-eb323db40e3b",
"created_at": "2026-08-19 10:08:41.189132+00:00",
"db_md5": "37bc6e3efd0710e8bf583f9e5bd36c49",
"submitter": "Ryan Teehan",
"github_login": "rteehas",
"run_id": "3ca168bb-73be-42fe-b985-993e9ba6b832",
"tier": "easy",
"dataset_id": "e4",
"status": "succee... | {
"score": {
"mean_loss": 7.278016479089935,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.005208333345750968
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 7.265719025421332,
"example_count": 1200... |
10547a96-cf53-47bc-965a-4899997b1458 | easy | erdavis0 | 2026-08-24 16:30:51.305981+00:00 | succeeded | b55593fc1495c8808062cd69554206bd2e9b09136f54bb7e28ab99edbd357bcd | 13,940 | null | """Field-aligned learned digit machine.
The fixed code in this model only exposes the public prompt grammar: delimiter
tokens select three fields and each field is right-aligned without interpreting
the value of any digit. All cross-field computation, local propagation,
workspace updates, and rendering are learned fr... | {
"id": "10547a96-cf53-47bc-965a-4899997b1458",
"created_at": "2026-08-24 16:30:51.305981+00:00",
"db_md5": "e47e69188792d13dc82454d0c65beec0",
"submitter": "Ethan",
"github_login": "erdavis0",
"run_id": "250cac5c-2d4d-4fa6-99b4-39ce32e2fc08",
"tier": "easy",
"dataset_id": "e5",
"status": "succeeded",... | {
"score": {
"mean_loss": 6.075356690461482,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.01375000002483527
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 6.016690106669883,
"example_count": 600,
... |
10589e56-0515-4b8c-bda9-7688bc28d6fd | easy | KaustubhKumar05 | 2026-08-25 18:18:12.099185+00:00 | succeeded | e61cc7591ec83d725aef1ff1fdd4693c311bba37127cc76ba9b0c3df8f54b631 | 16,286 | null | """conv_dist: conv_v2s_mom05 + a circular VALUE-DISTANCE loss.
Digit cross-entropy counts wrong symbols, which is anti-correlated with
numeric correctness on the commonest arithmetic error. Measured: for true
answer 120, CE scores 119 (off by ONE) at 8.00 and 220 (off by ONE HUNDRED)
at 4.00, because the borrow flips ... | {
"id": "10589e56-0515-4b8c-bda9-7688bc28d6fd",
"created_at": "2026-08-25 18:18:12.099185+00:00",
"db_md5": "fd430a49750f99034e68fb5a8169b1f4",
"submitter": "koz",
"github_login": "KaustubhKumar05",
"run_id": "dc1807a6-6e17-4b70-a156-08bb26c7c700",
"tier": "easy",
"dataset_id": "e6",
"status": "succee... | {
"score": {
"mean_loss": 6.038257837295532,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.2387387380003929
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 7.5551042556762695,
"example_count": 60,
... |
1059be8d-268c-4edd-b932-62c2d49b19e9 | easy | AdityaVG13 | 2026-08-15 09:16:20.123336+00:00 | succeeded | 738142651f7be34c15e82a9e7616f7f192e2a56988d45668213f3d4e46ec2584 | 109,666 | null | """AEGIS V9 v9_hopself: first-hop JOIN plus T=1 hop self-supervision in tape space.
One Layer Deeper submission candidate. All task transitions remain learned
end-to-end. The deterministic shell only identifies marker-delimited roles,
constructs a finite semantic tape, schedules learned interval programs, and
applies ... | {
"id": "1059be8d-268c-4edd-b932-62c2d49b19e9",
"created_at": "2026-08-15 09:16:20.123336+00:00",
"db_md5": "e4bd4d81f8942fc46b342978879ee683",
"submitter": "AdityaG",
"github_login": "AdityaVG13",
"run_id": "2b6ce53b-fc89-4807-83a3-b6ef39fa2a60",
"tier": "easy",
"dataset_id": "e1",
"status": "succeed... | {
"score": {
"mean_loss": 2.5763485431671143,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.6399999856948853
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 1.0822334289550781,
"example_count": 100,... |
10617ffc-2085-4206-9534-d09f829f494f | easy | 0Chris5R | 2026-08-07 09:28:02.623776+00:00 | succeeded | 006ac9bfb94f4c97c72c1390c599a09a7253e5f3be3acde9ddfa6d68b0d87fad | 12,681 | null | """Task-agnostic token-space function composer with exact structural T loops."""
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,
Tok... | {
"id": "10617ffc-2085-4206-9534-d09f829f494f",
"created_at": "2026-08-07 09:28:02.623776+00:00",
"db_md5": "cae065290f3371f84a8ee13be7328e80",
"submitter": "Chris ",
"github_login": "0Chris5R",
"run_id": "116de1dc-2570-4780-87a3-d33af56fa6e2",
"tier": "easy",
"dataset_id": "e5",
"status": "succeeded"... | {
"score": {
"mean_loss": 8.09032172374122,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.007083333333333334
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 8.61941525326716,
"example_count": 600,
... |
106dba76-2b63-4571-b782-ed400f4e0f3b | easy | pjjajal | 2026-08-07 03:29:05.627920+00:00 | succeeded | 1a46ee2c81feba82428d885122ae822f877fc4f68ad95a217d510ffcaf127502 | 12,473 | null | """Looped Transformer with an outer-product residual state.
Each token carries several residual streams. A dynamic read contracts those
streams to one Transformer-width vector; a dynamic write forms an outer product
between per-stream write weights and the Transformer output, then adds that
matrix to the identity-car... | {
"id": "106dba76-2b63-4571-b782-ed400f4e0f3b",
"created_at": "2026-08-07 03:29:05.627920+00:00",
"db_md5": "c5a74d6c37b769906545226159d466af",
"submitter": "pjjajal",
"github_login": "pjjajal",
"run_id": "d5e42e68-643a-489a-a6f6-166acb1e443c",
"tier": "easy",
"dataset_id": "e1",
"status": "succeeded"... | {
"score": {
"mean_loss": 2.0859556198120117,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.05833333171904087
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.014524459838867,
"example_count": 100,... |
107196aa-4682-4a32-bd64-2b2f6f223e5d | easy | shreyash-chonkie | 2026-08-26 14:58:53.728486+00:00 | failed | 2468677e32c1e002c6135f5e8413570bf43c895026bde9f29a87b636dea6f91e | 17,833 | null | """Online hard controller over rank-one associative memory updates."""
from __future__ import annotations
import math
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
assert_model_state,
)
P... | {
"id": "107196aa-4682-4a32-bd64-2b2f6f223e5d",
"created_at": "2026-08-26 14:58:53.728486+00:00",
"db_md5": "7509b2b3f470bbd5cdf2cc2a5c344d64",
"submitter": "Shreyash",
"github_login": "shreyash-chonkie",
"run_id": "21f370d9-7140-4c88-ad1c-58983cf4d09d",
"tier": "easy",
"dataset_id": "e1",
"status": "... | null
|
10755c8e-102e-4fd9-ac9b-c5be456b6f5d | easy | Gusanidas | 2026-08-21 09:22:14.378667+00:00 | succeeded | 7aaed32cb8d47c3757246eb2d6d07a8ccfdf9ec0dfc18409683d0bc38f2beb0c | 8,156 | null | """Fixed-N tiers as one learned transition composed T times.
`benchmark/runner.py:_scoring_split_names` drops both depth ladders, so the
score is the mean over `test` and `ood`. On a fixed-N dataset those two
splits reuse residues the training rows already touch (96% of `test` and 99%
of `ood` x-values on E1), so the... | {
"id": "10755c8e-102e-4fd9-ac9b-c5be456b6f5d",
"created_at": "2026-08-21 09:22:14.378667+00:00",
"db_md5": "a767142237e10554497dbec174d54b26",
"submitter": "Gusanidas",
"github_login": "Gusanidas",
"run_id": "e34ec5be-ed92-4831-abc1-adeb209c58a7",
"tier": "easy",
"dataset_id": "e1",
"status": "succee... | {
"score": {
"mean_loss": 0.9295993596315384,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.8416666686534882
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 0.25639811158180237,
"example_count": 100... |
1076dd94-d2e0-4fbe-bced-72b334fed6d1 | easy | sapient-sapiens | 2026-08-16 17:20:43.891995+00:00 | succeeded | 11fddf176f2313dd23f8b84a8b6f0d88a2d16e59bafd4f0d3f1af44237017c14 | 22,018 | null | """E6 generalization candidate: width-128, d_h=4, 6-block ACT64.
Architecture changes relative to the hosted ACT max-4 source:
- selectable recurrence depths are 1 through 64;
- canonical cumulative-mass halting genuinely removes halted examples from
later recurrent applications;
- the T-field representation preser... | {
"id": "1076dd94-d2e0-4fbe-bced-72b334fed6d1",
"created_at": "2026-08-16 17:20:43.891995+00:00",
"db_md5": "528d58393c19b1055302504017e45bc3",
"submitter": "Ertondy",
"github_login": "sapient-sapiens",
"run_id": "80aebeee-591f-43c5-88ac-fca8d02ca086",
"tier": "easy",
"dataset_id": "e6",
"status": "su... | {
"score": {
"mean_loss": 5.36680735300665,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.07364865311899702
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 4.851434230804443,
"example_count": 60,
... |
1078950b-db3d-4575-9c5a-0963ed1cc871 | easy | liam-gb | 2026-08-16 19:50:36.316929+00:00 | succeeded | fc98f09bb44e314df2633b71c4e733ee513887ef8e11317253f3e6e53490f7e3 | 16,365 | null | """rns_pm_gather_ct_widet_l32nore_fbmse: the grokking-regime probe, arm 2 of 2.
Same regime bundle as rns_pm_gather_ct_widet_l32nore_fb (full-batch 600 on E1,
weight decay 1.0 on tables) PLUS the loss the analytic-solution literature
derives under: mean-squared error between the softmax distribution and the
one-hot ta... | {
"id": "1078950b-db3d-4575-9c5a-0963ed1cc871",
"created_at": "2026-08-16 19:50:36.316929+00:00",
"db_md5": "9a2af27abc43620c69666f07f2117b02",
"submitter": "liam-gb",
"github_login": "liam-gb",
"run_id": "63a55c7d-15fb-448f-909d-aced9e507661",
"tier": "easy",
"dataset_id": "e1",
"status": "succeeded"... | {
"score": {
"mean_loss": 3.0571738481521606,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.25833333283662796
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 4.588258266448975,
"example_count": 100,... |
107fc7ac-6476-402e-b2ab-d0daaac75047 | easy | khushidahi | 2026-08-05 23:27:50.686681+00:00 | succeeded | 3b89766d7aaeb7caccce9c5b236647b899e8238d1148f6e3a364a308f573f416 | 19,855 | 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": "107fc7ac-6476-402e-b2ab-d0daaac75047",
"created_at": "2026-08-05 23:27:50.686681+00:00",
"db_md5": "a2382763ef8c2daefb58bc905e9ebc81",
"submitter": "khushidahi",
"github_login": "khushidahi",
"run_id": "e13bd873-b328-49e8-b5b1-7c2cf4129a47",
"tier": "easy",
"dataset_id": "e4",
"status": "succ... | {
"score": {
"mean_loss": 2.0779709815979004,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.008958333171904087
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.075486898422241,
"example_count": 120... |
10851a64-4a5f-46e0-8bb1-ffca1c64317a | easy | satyaki-up | 2026-08-07 17:06:14.586016+00:00 | succeeded | a4fc2dca02d6e8e7c4565f7e5018ac9340a43af6abc1ceac07303be1f97a8d4c | 11,908 | null | """Leaky LSTM with RNN optimization tricks from arXiv:1212.0901."""
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 = 2151
LAYE... | {
"id": "10851a64-4a5f-46e0-8bb1-ffca1c64317a",
"created_at": "2026-08-07 17:06:14.586016+00:00",
"db_md5": "c69bafd14189503b6b27deceafdaf490",
"submitter": "Satyaki",
"github_login": "satyaki-up",
"run_id": "15949404-56b7-4239-a1a4-f5ce8856b5e4",
"tier": "easy",
"dataset_id": "e5",
"status": "succeed... | {
"score": {
"mean_loss": 2.3310307973273936,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0033333333581686024
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.3296631366148124,
"example_count": 6... |
1085d947-6310-4983-a968-16828fe93730 | easy | pculaf | 2026-08-15 17:53:06.432493+00:00 | succeeded | 7f3b0e431a0672ac3d397e5004d0164899f69a4bb834319feaced0a56a99886c | 8,356 | 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": "1085d947-6310-4983-a968-16828fe93730",
"created_at": "2026-08-15 17:53:06.432493+00:00",
"db_md5": "8c9282b752a9f9568ffea905f2198dd3",
"submitter": "Pavle Culafic",
"github_login": "pculaf",
"run_id": "fcaab229-e35e-443d-8c80-ea1d909b00ac",
"tier": "easy",
"dataset_id": "e6",
"status": "succe... | {
"score": {
"mean_loss": 10.909693002700806,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.031981983222067356
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 15.003662109375,
"example_count": 60,
... |
108b0376-867f-4f5f-a78a-83235451a967 | easy | lzhao03 | 2026-08-17 16:21:52.789082+00:00 | succeeded | cdce2563421128c1c29769e2e71ef090ad26b3b7727e96d7b710426761946df2 | 14,716 | null | """Local-only categorical Markov bottleneck with an oracle recurrence count."""
from __future__ import annotations
from typing import cast
import torch
from torch import Tensor, nn
from torch.nn import functional as F
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
... | {
"id": "108b0376-867f-4f5f-a78a-83235451a967",
"created_at": "2026-08-17 16:21:52.789082+00:00",
"db_md5": "48a44db906a06e26248acf84046e54ad",
"submitter": "Leon Zhao",
"github_login": "lzhao03",
"run_id": "21d20ffa-b4f8-4826-a759-a54e499c5c0a",
"tier": "easy",
"dataset_id": "e1",
"status": "succeede... | {
"score": {
"mean_loss": 0.8677408024668694,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.8633333146572113
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 0.13949142396450043,
"example_count": 100... |
10978ea6-3b67-4e3c-8bb6-a3017c691ace | easy | oupadhyay | 2026-08-07 11:20:26.974349+00:00 | succeeded | 2b29ad99334078b8c072b263b1ec2f34ec61fc0126ef3e27dd0c590ddad12512 | 8,353 | null | """Complete compressed-axial forward path with the known-safe baseline optimizer."""
import torch
import torch.nn.functional as F
from torch import nn
from benchmark import OptimizerBundle, Submission, assert_model_state
PAD, N, X, T, ANS, DIGIT = 0, 2, 3, 4, 5, 7
MAX_STEPS, CHANNELS, STATE_ELEMENTS = 64, 64, 46634
... | {
"id": "10978ea6-3b67-4e3c-8bb6-a3017c691ace",
"created_at": "2026-08-07 11:20:26.974349+00:00",
"db_md5": "5d25b03810ff40e91e0e22adca08dd1b",
"submitter": "Ojasw Upadhyay",
"github_login": "oupadhyay",
"run_id": "9d10ba30-6165-437b-afaa-e19143bbdee5",
"tier": "easy",
"dataset_id": "e5",
"status": "s... | {
"score": {
"mean_loss": 2.4729713931222514,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.004583333345750968
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.486392658387599,
"example_count": 600... |
109dabda-260a-4533-83cb-11dcf6b8ed70 | easy | lpbb | 2026-08-12 08:45:19.590795+00:00 | succeeded | 318f76819438ea04f59ae695e7a430180975b571e93277e90538f76c51b0d060 | 25,156 | null | """DIGIT-STATE DISCRETE FLOW (v2) — diffusion-style quantized recurrence, e1.
v1 (submissions/agents/diffusion) put the recurrent state in a learned
codebook. Flaw found on paper: nothing ties step(enc(x)) to enc(f(x)), so
the code space can fragment into per-chain private paths and the coverage
argument (98-99% of e... | {
"id": "109dabda-260a-4533-83cb-11dcf6b8ed70",
"created_at": "2026-08-12 08:45:19.590795+00:00",
"db_md5": "19822823de3070d3f7ddbd9ffb211c18",
"submitter": "Lpbb",
"github_login": "lpbb",
"run_id": "79bb6c5c-f436-4a94-8c54-a7d2558ee0ab",
"tier": "easy",
"dataset_id": "e3",
"status": "succeeded",
"s... | {
"score": {
"mean_loss": 3.710428225266523,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.012500000004656613
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 3.7301933722281384,
"example_count": 800... |
10a424e8-8b12-4882-8d85-120d6c72750a | easy | yashkant | 2026-08-19 01:59:56.225168+00:00 | succeeded | 821067c8c528511c0db7e33e2118be1a26584442d85858058244d0f901aa97a1 | 41,440 | null | """Round2190 removes cross-step Dykstra correction carry.
The retained official carrier, primal recurrence, two learned projections,
antisymmetric set codes, parsed-T allocation, optimizer, and EMA are fixed.
The candidate reuses the learned initial A/B correction offsets on every
cycle instead of carrying hand-writte... | {
"id": "10a424e8-8b12-4882-8d85-120d6c72750a",
"created_at": "2026-08-19 01:59:56.225168+00:00",
"db_md5": "232687cd117a8d4468822e1ff2405cc0",
"submitter": "Yash Kant",
"github_login": "yashkant",
"run_id": "ed5bf457-2f3b-439c-b43f-40f59e38a1be",
"tier": "easy",
"dataset_id": "e4",
"status": "succeed... | {
"score": {
"mean_loss": 2.0719495177286724,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.010000000012417635
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.0744614062406677,
"example_count": 12... |
10a6c17c-39b7-43a5-a3d1-a2992b4a57d3 | easy | DDanlov | 2026-08-26 19:58:03.783619+00:00 | succeeded | b0909091203c3f45c5eb12ebb1529c3b28571ee31ba11dc1d99840ed866a8363 | 32,180 | 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": "10a6c17c-39b7-43a5-a3d1-a2992b4a57d3",
"created_at": "2026-08-26 19:58:03.783619+00:00",
"db_md5": "127bd99a12a5ac891d445825f9d96904",
"submitter": "DDanlov",
"github_login": "DDanlov",
"run_id": "4da31212-a71c-4524-a209-64a79beb3742",
"tier": "easy",
"dataset_id": "e5",
"status": "succeeded"... | {
"score": {
"mean_loss": 2.354898972738761,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0050000000248352685
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.343109138343366,
"example_count": 600... |
10acfa6f-0ee2-4cdb-964a-f17968dc8489 | easy | khushidahi | 2026-08-20 04:39:35.698402+00:00 | succeeded | d3feb43a02d0679ba7c9f86238efab152e8c491f8524064697599248b25b1416 | 33,914 | null | """R313: scalar-monomial semantic 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 reduction.
The transi... | {
"id": "10acfa6f-0ee2-4cdb-964a-f17968dc8489",
"created_at": "2026-08-20 04:39:35.698402+00:00",
"db_md5": "0d962631fbad7b9f595ee069c72db64e",
"submitter": "khushidahi",
"github_login": "khushidahi",
"run_id": "88d5ae76-1af6-422d-8862-a770a8243c54",
"tier": "easy",
"dataset_id": "e3",
"status": "succ... | {
"score": {
"mean_loss": 2.2384222994400025,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0068750000093132264
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.2202729486747104,
"example_count": 8... |
10b302c6-4594-422a-b8f8-b034ecdfb4c3 | easy | pculaf | 2026-08-25 14:04:16.272196+00:00 | succeeded | b3eb3fb8a08cc858183f73e7a647b0710c03cb193362f6b9cce0ea5e9c67ced0 | 8,288 | null | """Four-loop post-RMSNorm Transformer for One Layer Deeper."""
from __future__ import annotations
from collections.abc import Iterable
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
assert_m... | {
"id": "10b302c6-4594-422a-b8f8-b034ecdfb4c3",
"created_at": "2026-08-25 14:04:16.272196+00:00",
"db_md5": "06d3398262efcdfa4bf712724e235f4d",
"submitter": "Pavle Culafic",
"github_login": "pculaf",
"run_id": "842030fa-ff57-412e-ac94-58d4bb64abde",
"tier": "easy",
"dataset_id": "e3",
"status": "succe... | {
"score": {
"mean_loss": 2.1433198504185835,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.006250000004656613
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.1471044344213963,
"example_count": 80... |
10b80dc9-4827-49fe-9459-3a2e8c191e17 | easy | jordanrubin | 2026-08-13 18:11:59.897118+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": "10b80dc9-4827-49fe-9459-3a2e8c191e17",
"created_at": "2026-08-13 18:11:59.897118+00:00",
"db_md5": "3f710550312ea469d13b008d83980242",
"submitter": "Jordan Rubin",
"github_login": "jordanrubin",
"run_id": "806e9338-40f5-4237-92a8-e264b13334f5",
"tier": "easy",
"dataset_id": "e5",
"status": "s... | {
"score": {
"mean_loss": 2.851813155451894,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.028749998736505706
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.2380734835772236,
"example_count": 600... |
10c075f0-33c7-4892-a5c9-e4a8a2a1ad01 | easy | yashkant | 2026-08-12 04:35:42.380850+00:00 | succeeded | f7b6dd01606dc1c51f1060d09bf07ba00d9e15eea99e654f8339dc87ccb0cf0c | 31,388 | null | """Round1702 one-projection-per-T Dykstra recursion on the D416 carrier.
Each parsed public-T cycle applies one shared learned smooth projection and
alternates the A/B set code and correction memory across cycles. The immutable
anchor is reinjected on every call; all active cycle outputs are averaged, and
terminal CE... | {
"id": "10c075f0-33c7-4892-a5c9-e4a8a2a1ad01",
"created_at": "2026-08-12 04:35:42.380850+00:00",
"db_md5": "81d1341988474b2212473ffb3fa57d56",
"submitter": "Yash Kant",
"github_login": "yashkant",
"run_id": "0253105e-12bc-4cd7-9d56-b6a2cdf01a9c",
"tier": "easy",
"dataset_id": "e2",
"status": "succeed... | {
"score": {
"mean_loss": 4.1161879897117615,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.24916667863726616
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 6.408957481384277,
"example_count": 300,... |
10c17836-dd64-40be-bdc0-7dea8dd0a18a | easy | yunjiangster | 2026-08-26 23:29:40.674317+00:00 | succeeded | 66f744de53ceb62d0cf9b0d513f8b700a59795b7eabb2dc2c81fad7dd20c6d83 | 22,184 | null | """Quantized Markov Digit Machine for the One Layer Deeper benchmark.
The model parses the public prompt format into fixed-width decimal tapes, applies
one learned, tied transition once per requested time step, quantizes the residue
back to categorical digits after every step, and right-aligns the resulting
state with... | {
"id": "10c17836-dd64-40be-bdc0-7dea8dd0a18a",
"created_at": "2026-08-26 23:29:40.674317+00:00",
"db_md5": "b72da3b6cf37a861202a8d71e1f4ae72",
"submitter": "Yunjiang Jiang",
"github_login": "yunjiangster",
"run_id": "f5ea7a43-7215-4e04-b319-81a38ed734cb",
"tier": "easy",
"dataset_id": "e5",
"status":... | {
"score": {
"mean_loss": 4.377004721442717,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.012916666679084301
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 3.4214511078034815,
"example_count": 600... |
10c411b8-466b-484e-a110-18b815593d1d | easy | Yalyenea | 2026-08-11 21:34:25.148341+00:00 | succeeded | 6d78943e5f57f580ac7c177a353da4fb6a5993d04baab0720705b6f27c87e5b7 | 12,919 | null | """R4 full-bandwidth candidate with RoPE and a 2000-step cosine horizon."""
from __future__ import annotations
import math
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
assert_model_state,
... | {
"id": "10c411b8-466b-484e-a110-18b815593d1d",
"created_at": "2026-08-11 21:34:25.148341+00:00",
"db_md5": "e1ab4883f30066856bcfaaad405fd185",
"submitter": "yfff",
"github_login": "Yalyenea",
"run_id": "75b5851c-3120-4a1d-842e-54724dfb50cc",
"tier": "easy",
"dataset_id": "e5",
"status": "succeeded",
... | {
"score": {
"mean_loss": 6.557588841598314,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.009583333553746343
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 6.403753757476807,
"example_count": 600,... |
10c93d93-b78b-4f5b-a742-9adb8eab373a | easy | anirudh-chakravarthy | 2026-08-12 03:22:41.935398+00:00 | succeeded | 2fd94a6f05c47f68fecbb4decc6d5084b92622d0d201af4dd8dff730e189b206 | 14,573 | 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": "10c93d93-b78b-4f5b-a742-9adb8eab373a",
"created_at": "2026-08-12 03:22:41.935398+00:00",
"db_md5": "6d5f99ed163251ac6602ba8c4f8348bd",
"submitter": "Anirudh S Chakravarthy",
"github_login": "anirudh-chakravarthy",
"run_id": "e8220dfd-e0f8-404a-8fe1-e65c6eeb8aa6",
"tier": "easy",
"dataset_id": "... | {
"score": {
"mean_loss": 4.054627121515072,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.011250000074505806
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 4.962913134707463,
"example_count": 600,... |
10cd03f4-a06a-4b70-b999-a46b2564cbdf | easy | DDanlov | 2026-08-19 19:47:51.713458+00:00 | succeeded | a411f4eabffe555c2e818a6900a5978c3d42c5ce3e9bb076547d5226c3f66d0b | 17,727 | null | """
Dynamic Submission for One Layer Deeper API Benchmark
Config: dim=1952, num_heads=32, d_ff_mult=4, fixed_trec=112
"""
from __future__ import annotations
import math
from typing import Any, Optional, Tuple
import torch
import torch.nn as nn
import torch.nn.functional as F
from torch.optim import Optimizer
try:
... | {
"id": "10cd03f4-a06a-4b70-b999-a46b2564cbdf",
"created_at": "2026-08-19 19:47:51.713458+00:00",
"db_md5": "1af707a151e1a70bc7b3ba1e542205ea",
"submitter": "DDanlov",
"github_login": "DDanlov",
"run_id": "9558b082-aeea-4458-8136-004b23cc1658",
"tier": "easy",
"dataset_id": "e5",
"status": "succeeded"... | {
"score": {
"mean_loss": 2.4033051701503014,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.006666666679084301
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.479784481728558,
"example_count": 600... |
10ce6280-bc9b-4a83-82be-39e365b6a2ec | easy | lpbb | 2026-08-18 09:51:17.839668+00:00 | succeeded | 44c4fb64d99d9152aef2e7f067151e62e76170a9986cb45aab957aa2cb78eb2e | 4,457 | null | """S0 — control: same shape as S1 but 24 untied layers (fixed depth, no weight tying).
Measures the 'tying tax'; by construction its depth cannot follow T."""
from __future__ import annotations
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
ModelSpec,
Optimi... | {
"id": "10ce6280-bc9b-4a83-82be-39e365b6a2ec",
"created_at": "2026-08-18 09:51:17.839668+00:00",
"db_md5": "01eb56d416d674010c3e21901219e489",
"submitter": "Lpbb",
"github_login": "lpbb",
"run_id": "cc0bd537-9bc5-4705-b620-d4582c9306a4",
"tier": "easy",
"dataset_id": "e1",
"status": "succeeded",
"s... | {
"score": {
"mean_loss": 6.302286148071289,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.038333332166075706
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 6.250579357147217,
"example_count": 100,... |
10e01d1b-f992-4102-9cea-003f045036de | easy | repst | 2026-08-26 12:57:23.883828+00:00 | succeeded | 457f0bcab011dfebe0cc1cdf192171b098190fd4248ed63c77abeab2ee48981c | 125,355 | null | # R735: teacher-free complete-answer latent-program induction.
#
# Ground-truth answers enter only through ordinary complete-sequence likelihood.
# No label-derived behavior score, exact-match reward, elite program target, or
# detached program teacher is used by the active loss. A generic exact executor
# renders prop... | {
"id": "10e01d1b-f992-4102-9cea-003f045036de",
"created_at": "2026-08-26 12:57:23.883828+00:00",
"db_md5": "739196fabaa1ca350f7fd5e830405adf",
"submitter": "Ryan Epstein",
"github_login": "repst",
"run_id": "15669dfe-de0d-4297-aeb2-26cca15b6cc2",
"tier": "easy",
"dataset_id": "e1",
"status": "succeed... | {
"score": {
"mean_loss": 0.029658811283007524,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 1.0
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 0.029662268236279488,
"example_count": 100,
... |
10e1cf83-5bb6-4667-b14b-01d9adbf3a8e | easy | deepakvijaykee | 2026-08-10 14:21:37.617624+00:00 | succeeded | 2d73cab6962c65e112944048227af0b005c5d1771be6fd13c8596d797ab7901c | 21,069 | null | """One Layer Deeper candidate: tied digit-tape recurrence + verified deep supervision.
DESIGN
The task is T applications of an unknown map on Z_N, read and written as
decimal digit tokens. Five stages with explicit contracts; every free choice
lives in ExperimentConfig with its rationale.
parse input_ids -> Par... | {
"id": "10e1cf83-5bb6-4667-b14b-01d9adbf3a8e",
"created_at": "2026-08-10 14:21:37.617624+00:00",
"db_md5": "689e3091fa67adc822cd9d89c6e3cb55",
"submitter": "Deepak Vijaykeerthy",
"github_login": "deepakvijaykee",
"run_id": "0dac4869-e8d6-4565-be1d-42f5ff5e8d15",
"tier": "easy",
"dataset_id": "e3",
"s... | {
"score": {
"mean_loss": 2.4656638929471053,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.013125000018626452
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.4596510845219632,
"example_count": 80... |
10edb4ca-81af-4003-b25d-e2f4e701c955 | easy | oupadhyay | 2026-08-07 10:46:42.256308+00:00 | succeeded | f9ffbcd1770942e5c13bf44482d753915f0140dc9ac3bd54e37bbc5a542efa57 | 8,347 | null | """Complete compressed-axial forward path with the known-safe baseline optimizer."""
import torch
import torch.nn.functional as F
from torch import nn
from benchmark import OptimizerBundle, Submission, assert_model_state
PAD, N, X, T, ANS, DIGIT = 0, 2, 3, 4, 5, 7
MAX_STEPS, CHANNELS, STATE_ELEMENTS = 3, 32, 13162
... | {
"id": "10edb4ca-81af-4003-b25d-e2f4e701c955",
"created_at": "2026-08-07 10:46:42.256308+00:00",
"db_md5": "e120f71e2a797af4339cc9c8d071d1f3",
"submitter": "Ojasw Upadhyay",
"github_login": "oupadhyay",
"run_id": "e486e798-41ba-45ea-9144-949c338b635a",
"tier": "easy",
"dataset_id": "e5",
"status": "s... | {
"score": {
"mean_loss": 2.3253818339928314,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.00375
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.325507938861847,
"example_count": 600,
"... |
10eeacf5-f41d-4479-9287-b27930722b6c | easy | velocizapkar | 2026-08-16 07:30:14.688816+00:00 | succeeded | 7b7a7441ef93d0e13be3ddeae9abd377ebcf2268deeda1e97eeb5f263dd62b82 | 18,840 | null | """Field-relative learned-product workspace with a tied residual decoder.
Every contiguous digit run is treated by the same structural rule. Its
digits receive a learned field code and an LSD-zero place code. Independent
learned left/right projections form ordered elementwise products, and a
parameter-free route sum... | {
"id": "10eeacf5-f41d-4479-9287-b27930722b6c",
"created_at": "2026-08-16 07:30:14.688816+00:00",
"db_md5": "3c914bb29f013c760814ce6af7822599",
"submitter": "Aakanksh Zarapkar",
"github_login": "velocizapkar",
"run_id": "a89146c8-b9b3-49ef-b7b8-aba7607134e1",
"tier": "easy",
"dataset_id": "e6",
"statu... | {
"score": {
"mean_loss": 3.5192288160324097,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.10518018249422312
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.9798805713653564,
"example_count": 60,... |
10ef1fa3-1f9c-4c9e-9b9c-4a147decdd3b | easy | 0xDEADD06 | 2026-08-12 14:41:12.420627+00:00 | succeeded | c363e4de20aa8aae534c036860257f84eb2f617fd8b67960efacf5bd814695c8 | 5,741 | null | """ParaLSTM: recurrent LSTM with weight-tied passes for algorithmic tasks."""
from __future__ import annotations
import torch
import torch.nn.functional as F
from torch import nn, Tensor
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
assert_model_state,
)
D_MODEL... | {
"id": "10ef1fa3-1f9c-4c9e-9b9c-4a147decdd3b",
"created_at": "2026-08-12 14:41:12.420627+00:00",
"db_md5": "353e23904f87fc01bb2c2b696f6a40cc",
"submitter": "0xDEADD06",
"github_login": "0xDEADD06",
"run_id": "30a5f5c8-8210-4ec9-a404-8bcdc3de9ace",
"tier": "easy",
"dataset_id": "e5",
"status": "succee... | {
"score": {
"mean_loss": 2.5280397562007972,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.003750000111758709
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.6369191356036694,
"example_count": 60... |
10f136df-5d53-4e5d-aa3e-31c764be116a | easy | sapient-sapiens | 2026-08-04 16:05:24.148092+00:00 | succeeded | 6c7f97e576eff4be765ee10419618767e44af07975d03099909eec815a789028 | 18,010 | null | """Weight-tied recurrent-depth Transformer with progressive + hold losses.
Phase 2: prelude / recalled tied core / coda, differentiable hold gate,
fp32 core under disabled autocast, random train depth, K=72 at eval.
Extra depth readouts travel through the auxiliary return value into
token_training_loss. Wall-clock Ada... | {
"id": "10f136df-5d53-4e5d-aa3e-31c764be116a",
"created_at": "2026-08-04 16:05:24.148092+00:00",
"db_md5": "d5e2671ef7c8971a284322562d86109a",
"submitter": "Ertondy",
"github_login": "sapient-sapiens",
"run_id": "4f0107ce-895b-47c4-95b9-680bc3cd2dd4",
"tier": "easy",
"dataset_id": "e1",
"status": "su... | {
"score": {
"mean_loss": 2.32657479050859,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.051666667759418486
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.3939930191680565,
"example_count": 100,... |
10f29070-d45f-4d09-80f1-6c7d7097020c | easy | sapient-sapiens | 2026-08-09 15:30:19.533554+00:00 | succeeded | 08c425ce74467e65e89f51e6e7bd50f03b0ad70a98be34f91e3c39937b3a6b9e | 14,026 | null | """Near-500M adaptive recurrence grown from the hosted-working baseline."""
from __future__ import annotations
import math
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
TokenLossBatch,
... | {
"id": "10f29070-d45f-4d09-80f1-6c7d7097020c",
"created_at": "2026-08-09 15:30:19.533554+00:00",
"db_md5": "aa42bb2b6b1d97a88c7ef731109c6f79",
"submitter": "Ertondy",
"github_login": "sapient-sapiens",
"run_id": "03ae6bd5-0495-427d-b7d9-60738645d2d9",
"tier": "easy",
"dataset_id": "e1",
"status": "su... | {
"score": {
"mean_loss": 2.4424976180957003,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.006666666666666667
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.2684030617350963,
"example_count": 10... |
10f43a54-88de-4bf9-ae45-e2f2ad9609d5 | easy | AFeinfield | 2026-08-12 23:39:51.351674+00:00 | succeeded | b01f9887063cf510a7495631f6ffd2b40db80e2ae03285caee33c789a45b52c2 | 4,661 | null | """Looped (weight-tied) bidirectional Transformer, d=256, 8 iterations.
Idea #1 from ideas.md: one shared block iterated K times = serial computation
prior. Same parameter count order as a 1-block model; depth comes from reuse.
Warmup + cosine LR. No compile (60s budget).
"""
from __future__ import annotations
impor... | {
"id": "10f43a54-88de-4bf9-ae45-e2f2ad9609d5",
"created_at": "2026-08-12 23:39:51.351674+00:00",
"db_md5": "5ad2bd0935d862998b43da0a23c2d284",
"submitter": "AFeinfield",
"github_login": "AFeinfield",
"run_id": "3b032204-1adc-4e8d-936c-a41734095805",
"tier": "easy",
"dataset_id": "e1",
"status": "succ... | {
"score": {
"mean_loss": 2.885236620903015,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0033333334140479565
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.774782419204712,
"example_count": 100... |
10f47c7e-2d23-4ed9-ac79-9307599689df | easy | sapient-sapiens | 2026-08-14 18:39:23.679021+00:00 | succeeded | 28ba469aae3e42ca679270105f219b6bd47433fa80f83dc154b3c3e55e44c7c3 | 14,299 | null | """Digit-space recurrent operator for repeated modular squaring.
Nothing here converts a token into a decimal number. Digits stay as embeddings,
so what a digit is worth, and every carry and reduction, is learned. That also
sidesteps a hard numeric wall: at Hard's modulus scale `x^2` runs past the 53-bit
exact-integer... | {
"id": "10f47c7e-2d23-4ed9-ac79-9307599689df",
"created_at": "2026-08-14 18:39:23.679021+00:00",
"db_md5": "1efadda9e01ae84ec2857432a666c542",
"submitter": "Ertondy",
"github_login": "sapient-sapiens",
"run_id": "6070d419-554b-4e76-a28e-ae11fe5f2538",
"tier": "easy",
"dataset_id": "e4",
"status": "su... | {
"score": {
"mean_loss": 2.0639806913442635,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.006562500000000001
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.0661830709060154,
"example_count": 12... |
1105c965-33ef-4f52-9557-cb88b47d9ff6 | easy | oupadhyay | 2026-08-12 06:53:16.974841+00:00 | succeeded | 1cd48196fb1baec7ca14734984251cc6fc215105293148a7825a20e07e11920d | 8,245 | null | """Standalone RMS/axial crossing C44 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": "1105c965-33ef-4f52-9557-cb88b47d9ff6",
"created_at": "2026-08-12 06:53:16.974841+00:00",
"db_md5": "2b94164ed576ad48f635ec25eac875cb",
"submitter": "Ojasw Upadhyay",
"github_login": "oupadhyay",
"run_id": "4fa9661c-1e47-4f8b-bab2-8e76df9afd3d",
"tier": "easy",
"dataset_id": "e5",
"status": "s... | {
"score": {
"mean_loss": 2.184664904524949,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.006666666666666667
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.147030804723902,
"example_count": 600,... |
110afd23-b478-4d7d-b854-19fbe88b42cb | easy | khushidahi | 2026-08-07 02:59:24.930222+00:00 | succeeded | e2d420cec1320af5fa45f07025eb35fa2e6194f570c289c590ba880ad610efb6 | 18,223 | 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": "110afd23-b478-4d7d-b854-19fbe88b42cb",
"created_at": "2026-08-07 02:59:24.930222+00:00",
"db_md5": "27fccdf66ef394d19de050330284cb73",
"submitter": "khushidahi",
"github_login": "khushidahi",
"run_id": "b1b61bd9-0ef9-45f8-8b30-aba49ce2fde2",
"tier": "easy",
"dataset_id": "e5",
"status": "succ... | {
"score": {
"mean_loss": 2.1588693857192993,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.006250000325962901
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.156792402267456,
"example_count": 600... |
1116fa53-c498-463b-a3d4-8056053baae8 | easy | sapient-sapiens | 2026-08-20 12:14:42.492459+00:00 | succeeded | f5467ffe8b0d84e2e2c145a9d53fcebd801be2e3074d46ca691853898c419e5d | 33,891 | 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": "1116fa53-c498-463b-a3d4-8056053baae8",
"created_at": "2026-08-20 12:14:42.492459+00:00",
"db_md5": "390c269976813008f48424275109c09f",
"submitter": "Ertondy",
"github_login": "sapient-sapiens",
"run_id": "0cf7e321-ee21-45eb-bb60-f0e253042c87",
"tier": "easy",
"dataset_id": "e5",
"status": "su... | {
"score": {
"mean_loss": 2.2081682141666334,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.01166666670391957
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.211505325252165,
"example_count": 600,... |
1130cba8-00af-4bf4-9dcb-fa2250f82b7c | easy | khushidahi | 2026-08-09 04:57:50.180949+00:00 | succeeded | eaece6c41c6fbc168171cfa99f3d5db1a094cc87cab44547c4e0ad7a78535a47 | 23,845 | 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": "1130cba8-00af-4bf4-9dcb-fa2250f82b7c",
"created_at": "2026-08-09 04:57:50.180949+00:00",
"db_md5": "f7b43a5876d9b346016283b8892b474b",
"submitter": "khushidahi",
"github_login": "khushidahi",
"run_id": "1a3857ab-6809-4032-a6c5-f4a27f0d535d",
"tier": "easy",
"dataset_id": "e5",
"status": "succ... | {
"score": {
"mean_loss": 2.1956360300180613,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.010000000024835268
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.1946935653686523,
"example_count": 60... |
11432b52-7145-4229-98bb-c64c4506c887 | easy | erdavis0 | 2026-08-23 12:56:31.288127+00:00 | succeeded | 40c19379d896a602048fe8a8415187772a6da9fa3fd51caad85c78fe8890d14c | 21,871 | null | """Learned field-output Neural GPU on a generic rectangular grid.
Public delimiters route opaque categories into separate right-aligned N, X,
and T tapes. Complete X/T input-place tapes are broadcast across every output
column of a rectangular computational fabric. Three persistent interaction,
carry, and modulus gr... | {
"id": "11432b52-7145-4229-98bb-c64c4506c887",
"created_at": "2026-08-23 12:56:31.288127+00:00",
"db_md5": "452f0b4d320ad628e99264c2a524a05c",
"submitter": "Ethan",
"github_login": "erdavis0",
"run_id": "951ae2e1-58b9-49d7-88e5-be7e519258a4",
"tier": "easy",
"dataset_id": "e5",
"status": "succeeded",... | {
"score": {
"mean_loss": 2.1495579383327916,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0037500000248352687
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.151994606839167,
"example_count": 60... |
114f5d4d-6b0e-43c9-b9d3-bedb8466530e | easy | DDanlov | 2026-08-28 22:06:44.867261+00:00 | succeeded | 425640900914ba413eb476092ab0f97d8d9c155f291f6c8af42e760a9d90fcbd | 15,069 | null | """
Trial 32: Model C: D=768, H=12, M=4, u=2 (T=8)
"""
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": "114f5d4d-6b0e-43c9-b9d3-bedb8466530e",
"created_at": "2026-08-28 22:06:44.867261+00:00",
"db_md5": "04e53b654b83c1c6bdadf31109915fd7",
"submitter": "DDanlov",
"github_login": "DDanlov",
"run_id": "38c0a7b9-4842-4eae-a509-12b121162a2e",
"tier": "easy",
"dataset_id": "e5",
"status": "succeeded"... | {
"score": {
"mean_loss": 179.03715318940564,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0020833333457509675
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 177.00336210823914,
"example_count": 6... |
1151a13a-dc8b-4220-a949-d20e4beb013c | easy | DDanlov | 2026-08-08 08:58:06.232131+00:00 | failed | 0f2b5891ac7e06e7be96ad6edae556268ff92c6bea9f5ac9d0a1e2762ab4a19c | 8,653 | 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
VOCAB_SIZE = 17
DIGIT_OFFSET = 7
class ModelConfig:
def __init__(self, vocab_size: int, max_seq_len: int):
self.vocab_size = vocab_size
self.max_seq_le... | {
"id": "1151a13a-dc8b-4220-a949-d20e4beb013c",
"created_at": "2026-08-08 08:58:06.232131+00:00",
"db_md5": "4b5ef54b693d50c480e30508e164d0ec",
"submitter": "DDanlov",
"github_login": "DDanlov",
"run_id": "d9d484d0-3236-4d53-97ad-558e4f03350a",
"tier": "easy",
"dataset_id": "e1",
"status": "failed",
... | null
|
1154fc9d-4c08-4d77-b55b-bb0f750fa6b9 | easy | erdavis0 | 2026-08-19 08:54:10.519415+00:00 | succeeded | 7bf30e19d90e7f6cadb4ad67d0b2cd51634a4fd2d29a507be6f840a0dd78f8c1 | 3,628 | null | """Basic single-pass Transformer with PyTorch AdamW."""
from __future__ import annotations
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
assert_model_state,
)
D_MODEL = 128
NUM_HEADS = 4
... | {
"id": "1154fc9d-4c08-4d77-b55b-bb0f750fa6b9",
"created_at": "2026-08-19 08:54:10.519415+00:00",
"db_md5": "eaac69fe345f82973fcf392c504a40ff",
"submitter": "Ethan",
"github_login": "erdavis0",
"run_id": "256b492f-461e-4e42-9e37-d4d77636aba9",
"tier": "easy",
"dataset_id": "e7",
"status": "succeeded",... | {
"score": {
"mean_loss": 4.0618016719818115,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.02103386865928769
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 5.473084926605225,
"example_count": 85,
... |
11576073-5e1b-4465-a7a9-7a646661c246 | easy | anirudh-chakravarthy | 2026-08-12 18:45:39.121793+00:00 | succeeded | 5ddfc4e47ac9a21a687d60fed97184e54f488c5affed7e2352521d24db47d221 | 11,743 | null | """A diagonal Neural GPU over binary-coded decimal number cells."""
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": "11576073-5e1b-4465-a7a9-7a646661c246",
"created_at": "2026-08-12 18:45:39.121793+00:00",
"db_md5": "c23265b0d992cc3718e608ef72fd5515",
"submitter": "Anirudh S Chakravarthy",
"github_login": "anirudh-chakravarthy",
"run_id": "c346da03-de33-4136-9296-775c1d9195d4",
"tier": "easy",
"dataset_id": "... | {
"score": {
"mean_loss": 2.1536510466773193,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.012916666691501936
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.153927645875734,
"example_count": 600... |
115bbdce-05ec-4d7e-8cca-57f9aee6e4e2 | easy | epicosy | 2026-08-22 23:03:31.451346+00:00 | succeeded | b44428c2eb050922fc23594e21180e85d23fcd3ac6f914066b89fd966b9a12e6 | 5,430 | null | """Latent-state Universal Transformer baseline for the Easy tier.
The prompt stream is encoded once and kept fixed. A single shared block
recurrently updates learned latent/query tokens through cross-attention to the
prompt. The final logits remain aligned with the original input sequence, as
required by the public su... | {
"id": "115bbdce-05ec-4d7e-8cca-57f9aee6e4e2",
"created_at": "2026-08-22 23:03:31.451346+00:00",
"db_md5": "b553ea8294865682d9c0b34d689a0a79",
"submitter": "Eduard Pinconschi",
"github_login": "epicosy",
"run_id": "5f39904a-9fdf-4244-a217-35201710dc49",
"tier": "easy",
"dataset_id": "e6",
"status": "... | {
"score": {
"mean_loss": 2.361065149307251,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.02342342445626855
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.379232883453369,
"example_count": 60,
... |
115f0b55-e595-49be-bf82-59dedfffca1b | easy | DDanlov | 2026-08-26 20:49:44.889296+00:00 | succeeded | 83da4e39f226e5437784047a0519f8826743b1e828860b9252b2b48c04451325 | 34,346 | 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": "115f0b55-e595-49be-bf82-59dedfffca1b",
"created_at": "2026-08-26 20:49:44.889296+00:00",
"db_md5": "50b60d7add1e01adc3cd02382c837e3f",
"submitter": "DDanlov",
"github_login": "DDanlov",
"run_id": "c7e091d8-ffed-42a3-af4c-c2040843274b",
"tier": "easy",
"dataset_id": "e5",
"status": "succeeded"... | {
"score": {
"mean_loss": 2.4302413690067244,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0004166666666666667
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.437003049080682,
"example_count": 60... |
11651f46-ac42-458b-a4a2-39e3b6ee01c8 | easy | himalalps | 2026-08-10 01:44:00.984575+00:00 | succeeded | a70b575a7d687dcc965e19809393d2da11d6157062f41c923941db8cc9159292 | 85,615 | 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": "11651f46-ac42-458b-a4a2-39e3b6ee01c8",
"created_at": "2026-08-10 01:44:00.984575+00:00",
"db_md5": "90dc1e6ba9b857735d5bab8d8e7ed789",
"submitter": "Haoyu Tang",
"github_login": "himalalps",
"run_id": "460cc15f-35ef-48be-95a5-9e6b86383e7e",
"tier": "easy",
"dataset_id": "e5",
"status": "succe... | {
"score": {
"mean_loss": 2.1840245341462596,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.010833333233992258
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.1612017154693604,
"example_count": 60... |
11657475-bc02-40d7-b5f7-b8d354f1a7e0 | easy | 0Chris5R | 2026-08-01 07:17:08.115801+00:00 | succeeded | c07e842ca4bf659e0645f28ba91d449e97f49f1cd85ab49cf0ce902b13209b50 | 8,074 | null | """Learned convolutional-GRU over an aligned field-by-digit workspace.
The public field markers and decimal significance define only the layout of
the neural workspace. Every transition and output is learned; there is no
arithmetic routine, recurrence assumption, lookup table, or generated target.
"""
from __future__... | {
"id": "11657475-bc02-40d7-b5f7-b8d354f1a7e0",
"created_at": "2026-08-01 07:17:08.115801+00:00",
"db_md5": "e63df199be8694d35bedac6a389e3c01",
"submitter": "Chris ",
"github_login": "0Chris5R",
"run_id": "6bc79d96-b0d1-4b49-b2ad-2727d8f2ef74",
"tier": "easy",
"dataset_id": "e5",
"status": "succeeded"... | {
"score": {
"mean_loss": 8.587629980281577,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.004583333345750968
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 8.460065148871042,
"example_count": 600,... |
116587bd-0a58-4e2f-847c-df8d8e10f420 | easy | yashkant | 2026-08-14 17:20:54.505983+00:00 | succeeded | 2c26d536fd9f60639c2616124d74897a6b0b2b0cc5f4f11a529d4e6f656433ca | 34,476 | null | """Round1702 one-projection-per-T Dykstra recursion on the D416 carrier.
Each parsed public-T cycle applies one shared learned smooth projection and
alternates the A/B set code and correction memory across cycles. The immutable
anchor is reinjected on every call; all active cycle outputs are averaged, and
terminal CE... | {
"id": "116587bd-0a58-4e2f-847c-df8d8e10f420",
"created_at": "2026-08-14 17:20:54.505983+00:00",
"db_md5": "07e5c5b1cf1bd931e308fd65cba96476",
"submitter": "Yash Kant",
"github_login": "yashkant",
"run_id": "11b5f738-8aad-4a15-8aa8-571f8058e912",
"tier": "easy",
"dataset_id": "e2",
"status": "succeed... | {
"score": {
"mean_loss": 4.421582102775574,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.29500002786517143
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 7.136323928833008,
"example_count": 300,
... |
11678e0b-ffd0-422b-b76c-3d97b81d5d77 | easy | Yalyenea | 2026-08-11 19:37:40.424807+00:00 | succeeded | 7da267286413b58fe0f8c17400eee71725caae63b5a7b75de88cd7942e4f417d | 12,109 | null | """R4 full-bandwidth temporal-feedback candidate with a 2000-step cosine horizon."""
from __future__ import annotations
import math
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
assert_mode... | {
"id": "11678e0b-ffd0-422b-b76c-3d97b81d5d77",
"created_at": "2026-08-11 19:37:40.424807+00:00",
"db_md5": "6266ea0ff31a4aeff9fc9663266e3673",
"submitter": "yfff",
"github_login": "Yalyenea",
"run_id": "e4a92b95-8f44-430b-a7a5-621995824027",
"tier": "easy",
"dataset_id": "e5",
"status": "succeeded",
... | {
"score": {
"mean_loss": 6.271699144765064,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.015833334252238273
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 6.066876411437988,
"example_count": 600,... |
116abb62-23ff-4102-b1d4-845e971a665b | easy | DDanlov | 2026-08-10 09:43:22.694409+00:00 | succeeded | 28f7b5a8e08ce69fa9fd8dc2dd9f6c496232235f4b313af0d87adb6869292b8a | 15,745 | null | from __future__ import annotations
import sys
import os
import math
import time
import torch
import torch.nn as nn
import torch.nn.functional as F
from torch.optim.optimizer import Optimizer
VOCAB_SIZE = 17
DIGIT_OFFSET = 7
def init_cutting_the_skip(module):
if isinstance(module, nn.Linear):
nn.init.ortho... | {
"id": "116abb62-23ff-4102-b1d4-845e971a665b",
"created_at": "2026-08-10 09:43:22.694409+00:00",
"db_md5": "0b55e4a9bbf44ecd9df178315317af5b",
"submitter": "DDanlov",
"github_login": "DDanlov",
"run_id": "369f976a-2939-4a20-abdb-1a837587408c",
"tier": "easy",
"dataset_id": "e5",
"status": "succeeded"... | {
"score": {
"mean_loss": 2.545136763065793,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.0141666666790843
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 2.5469366322718394,
"example_count": 600,
... |
116d65de-b187-4caf-883a-30d09c0e910d | easy | isaac0804 | 2026-08-06 14:38:32.481433+00:00 | succeeded | b2f8a6c35bc90b0ee273cb44e9c904b62b0011e27a5cfd7a2efaa7a534b5ceba | 9,324 | null | """Diagnostic-only: isolate whether ONE of E3's own moduli can be learned to
held-out generalization at all, independent of the 26-other-moduli joint
training question.
docs/findings.md ("Session log: numerically diagnosing the E3 collapse")
established that E3's held-out failure is not step-starvation or capacity --
... | {
"id": "116d65de-b187-4caf-883a-30d09c0e910d",
"created_at": "2026-08-06 14:38:32.481433+00:00",
"db_md5": "4dc9c05312e8414951f74572b9af038c",
"submitter": "Isaac Yong",
"github_login": "isaac0804",
"run_id": "210e35b5-f8fc-4085-a7c5-fa9f25980123",
"tier": "easy",
"dataset_id": "e3",
"status": "succe... | {
"score": {
"mean_loss": 14.3477088147147,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.008125000009313226
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 14.487510529094049,
"example_count": 800,... |
116d72ef-d99c-4ce1-9ab7-137115a9492f | easy | neel04 | 2026-08-11 13:07:43.598423+00:00 | failed | ea4f6391b3cd377c97a17de3493d9236734def90592e181cfc805b11a5837492 | 17,362 | null | """Layer 6: stationary NAG with tied decoding and an auxiliary critic."""
from __future__ import annotations
import math
import torch
import torch.nn.functional as F
from torch import Tensor, nn
from benchmark import (
BatchReuseContext,
ModelSpec,
OptimizerBundle,
OptimizerSpec,
Submission,
... | {
"id": "116d72ef-d99c-4ce1-9ab7-137115a9492f",
"created_at": "2026-08-11 13:07:43.598423+00:00",
"db_md5": "dd890b6d242e892d1c71f2b2860aa360",
"submitter": "Neel Gupta",
"github_login": "neel04",
"run_id": "aebe3e11-f97c-492d-8ab5-a93b3d2de032",
"tier": "easy",
"dataset_id": "e1",
"status": "failed",... | null
|
1171dac2-c9ed-46bd-851d-314bb1bf9893 | easy | chuk-uzowihe | 2026-08-10 06:10:01.401024+00:00 | succeeded | 8b3ee3dfedeccb7b026a7c04e55c9b71cf9ec7e5c6c27e5803de01574b98a2da | 64,260 | null | """One Layer Deeper submission: depth-routed MoEUT over a 2D digit lattice.
v9: the recurrent block is a single MoEUT layer (refs/moeut) — SwitchHead
attention + sigma-MoE feedforward — whose routers are conditioned on the
loop depth (Fourier features of the iteration index) so one shared bank can
express phase-depend... | {
"id": "1171dac2-c9ed-46bd-851d-314bb1bf9893",
"created_at": "2026-08-10 06:10:01.401024+00:00",
"db_md5": "6b17fa48d6e36f64317b0ba6e648b045",
"submitter": "Chuk Uzowihe",
"github_login": "chuk-uzowihe",
"run_id": "c2958be5-f41e-4bd3-800f-bb4662007bd7",
"tier": "easy",
"dataset_id": "e1",
"status": "... | {
"score": {
"mean_loss": 10.542397499084473,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.054999999701976776
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 10.68594741821289,
"example_count": 100... |
117502a3-40f1-4ccd-bb6b-9a34f404dcc2 | easy | yashkant | 2026-08-13 00:33:04.398921+00:00 | succeeded | fcb6e4f4719005cd2391ce4c5fe2e3637adbe575d6494caf8c70cbf0ad8d441c | 35,038 | null | """Round1542 low-capacity periodic-fold answer/latent state machine.
A width-32 answer/latent pair is refined by one shared hidden-16 MLP once per
public-T macro. The candidate advances each field through a differentiable
triangle-wave fold, giving bounded carry-free wraparound; the equal-work
control uses a conventio... | {
"id": "117502a3-40f1-4ccd-bb6b-9a34f404dcc2",
"created_at": "2026-08-13 00:33:04.398921+00:00",
"db_md5": "e6e641a83d0464513d5fee5a23fb1f78",
"submitter": "Yash Kant",
"github_login": "yashkant",
"run_id": "608787ea-2048-4b2c-9241-93b405657050",
"tier": "easy",
"dataset_id": "e1",
"status": "succeed... | {
"score": {
"mean_loss": 5.0914623737335205,
"primary_metric": "mean_exact_accuracy",
"num_measurements": 2,
"mean_exact_accuracy": 0.06333333253860474
},
"seeds": [
{
"seed": 74,
"evaluation": {
"ood": {
"loss": 5.962334156036377,
"example_count": 100,... |
117809b7-c7bf-4d73-9a2f-fc9ebad4c912 | easy | DDanlov | 2026-08-13 22:43:04.098633+00:00 | failed | c1881ee0e526a2c6b0d05a3c0df370f7ecde20e4bc60ea90c51a5490a1bbffff | 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": "117809b7-c7bf-4d73-9a2f-fc9ebad4c912",
"created_at": "2026-08-13 22:43:04.098633+00:00",
"db_md5": "8145d61d5cc8fefe6352c534ed6b2272",
"submitter": "DDanlov",
"github_login": "DDanlov",
"run_id": "c419171b-9fc7-4f81-aee1-38aa8ad6f953",
"tier": "easy",
"dataset_id": "e3",
"status": "failed",
... | null
|
1178b9e1-a662-42c1-91f7-652219513782 | easy | DDanlov | 2026-08-11 19:30:39.633752+00:00 | failed | ece22607915a3121b3cac12e5e878176029e07bde098d7921fcb37551012e19e | 19,000 | null | from __future__ import annotations
import sys
import os
import math
import time
import torch
import torch.nn as nn
import torch.nn.functional as F
from torch.optim.optimizer import Optimizer
VOCAB_SIZE = 17
DIGIT_OFFSET = 7
class RohanShampoo(Optimizer):
"""Self-contained RohanShampoo optimizer with bias-correcte... | {
"id": "1178b9e1-a662-42c1-91f7-652219513782",
"created_at": "2026-08-11 19:30:39.633752+00:00",
"db_md5": "f4e23ee1cfc8a2d0f11d9da8e9206567",
"submitter": "DDanlov",
"github_login": "DDanlov",
"run_id": "9e3117bf-2c88-4fea-bdfc-1d927822dafa",
"tier": "easy",
"dataset_id": "e1",
"status": "failed",
... | null
|
117b55cd-4b73-4b32-a82d-ee659981fdc9 | easy | DDanlov | 2026-08-23 22:20:48.601489+00:00 | failed | ac03df670531977fd0903ca52f178b1ab3d0a6812a908395020fc347ac53bc7d | 32,662 | null | """
Unified Hybrid Masked Deliberation Model (40.4M Parameters) - Pure Skipless, No Cross-Attention
Architecture:
- Scale: dim=1536, num_heads=24, d_ff=6144 (Exactly 40,437,888 parameters)
- Hybrid Prefix-Bidirectional + Causal Target Attention Mask
- Token Layout: [Prompt (L_p), 4 Pause Registers, Target Tokens (L_tgt... | {
"id": "117b55cd-4b73-4b32-a82d-ee659981fdc9",
"created_at": "2026-08-23 22:20:48.601489+00:00",
"db_md5": "bbf26c9717a8a0b8575e7854f359678d",
"submitter": "DDanlov",
"github_login": "DDanlov",
"run_id": "d6e96290-bb0f-4c14-91b6-297132e49ddc",
"tier": "easy",
"dataset_id": "e5",
"status": "failed",
... | null
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.