SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2110.02037 · ICLR · 2022

Autoregressive Diffusion Models

Tim Salimans, Ben Poole, Emiel Hoogeboom, Rianne Van Den Berg, Alexey Gritsenko, Jasmijn Bastings

arXiv · PDF · Open in the Atlas

Code that ran

2 of 2 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (1 of them constructed an object rather than computing a result; 0 honoured the contract we drafted, 0 violated it and 1 has no contract verdict; 1 failure of our instrument; 0 no recorded run), measured 2026-11-22; repository links from the Papers with Code archive attributed under CC BY-SA, no affiliation; bodies inline only under a permissive licence as GitHub detects it. Not a reproduction of the paper's results.

"Ran" means the function executed on a synthesized input and returned a value. It is not a reproduction of the paper's results. This record, dated, as JSON: /api/ran/2110.02037.json · badge: /api/ran/2110.02037.svg

Each repository's role is a label: a classifier's output recorded on the link, or, where no role is recorded, what the Papers with Code archive marked. The line under it names its basis. Neither says who wrote the code.

RepositoryRoleRan
princetonlips/mam community
Papers with Code archive
1 ran · 0 no recorded run
microsoft/evodiff community
Papers with Code archive
1 ran · 0 no recorded run
FunctionStatusWhere it lives
MAM Ran
ran on a synthesized input · constructed an object; no result was computed
princetonlips/mam/mam/models/mam_eb.py
pointer only (licence: NONE) · get_code("74287fe6db7d5ad1")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, princetonlips/mam when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

generate_oaardm_order_opt Ran
ran; our contract draft was wrong, not the code
microsoft/evodiff/evodiff/generate.py
code served (permissive licence) · get_code("d1d0667ba9b2757b")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1) and returned tuple. The image it ran in was not recorded. The code mentions cuda, and the harness aliases cuda to the CPU for every sample; whether this run needed that is not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import torch.nn as nn
import numpy as np

torch.manual_seed(0)

# Build a minimal tokenizer that has the required attributes and methods
class MinimalTokenizer:
    def __init__(self):
        self.all_aas = ['A', 'C', 'D', 'E', 'F']  # 5 amino acids
        self.mask_id = 0
    
    def untokenize(self, s):
        # Convert tensor of indices to string
        return ''.join([self.all_aas[int(idx)] if int(idx) < len(self.all_aas) else 'X' for idx in s])

# Build a minimal model that satisfies how it's used in the function
class MinimalModel(nn.Module):
    def __init__(self, seq_len, num_aas):
        super().__init__()
        self.seq_len = seq_len
        self.num_aas = num_aas
        # Simple linear layer that outputs logits for each position and amino acid
        self.linear = nn.Linear(seq_len, seq_len * num_aas)
    
    def forward(self, sample, timestep):
        # sample: (batch_size, seq_len)
        # timestep: (batch_size,)
        # output: (batch_size, seq_len, num_aas)
        batch_size = sample.shape[0]
        # Flatten sample and pass through linear layer
        x = sample.float()
        out = self.linear(x)
        # Reshape to (batch_size, seq_len, num_aas)
        out = out.view(batch_size, self.seq_len, self.num_aas)
        return out

# Create instances
seq_len = 4
batch_size = 2
num_aas = 5

tokenizer = MinimalTokenizer()
model = MinimalModel(seq_len=seq_len, num_aas=num_aas)

ARGS = {
    'model': model,
    'tokenizer': tokenizer,
    'seq_len': seq_len,
    'penalty': None,
    'batch_size': batch_size,
    'device': 'cpu'
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
Imports
numpy, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type tuple — a tuple output has no array shape of its own; none is stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block

Also present in other papers' code

Of 2 samples lifted from this paper's linked repositories, 2 carry a function identity and 0 do not. Of the 2 distinct functions with an identity, 0 are also present in other papers whose code we hold (0 with at least one paper dated earlier, 0 the same month, 0 later) and 2 are only here among the 52,977 papers whose code Syntology holds (as of 2026-09-26).

Only here means: among the 52,977 papers whose code Syntology holds (51,887 with lifted samples, 1,090 through their assemblies only), as of 2026-09-26, no other paper's lifted samples or current assembly bricks contain an identical function body after normalisation.

Presence, not direction: an identical function body in two papers' repositories does not say who wrote it, which came from which, or that either paper's method rests on it. Matching is exact after normalisation, so a renamed or edited function is not matched. Dates are when papers appeared, not when code was written. Only .py files the harvester lifted are covered; notebooks are not read. A new-style arXiv id YYMM.NNNN dates a paper to the month 20YY-MM; otherwise Paper.year dates it to the year; otherwise it is undated. Dates are when the paper appeared, not when the code was written. Read 2026-11-22.

Repositories linked to this paper

Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.

Abstract

We introduce Autoregressive Diffusion Models (ARDMs), a model class encompassing and generalizing order-agnostic autoregressive models (Uria et al., 2014) and absorbing discrete diffusion (Austin et al., 2021), which we show are special cases of ARDMs under mild assumptions. ARDMs are simple to implement and easy to train. Unlike standard ARMs, they do not require causal masking of model representations, and can be trained using an efficient objective similar to modern probabilistic diffusion models that scales favourably to highly-dimensional data. At test time, ARDMs support parallel generation which can be adapted to fit any given generation budget. We find that ARDMs require significantly fewer steps than discrete diffusion models to attain the same performance. Finally, we apply ARDMs to lossless compression, and show that they are uniquely suited to this task. Contrary to existing approaches based on bits-back coding, ARDMs obtain compelling results not only on complete datasets, but also on compressing single data points. Moreover, this can be done using a modest number of network calls for (de)compression due to the model's adaptable parallel generation.

For agents

The same record, over MCP at https://syntology.ai/mcp:

get_harvested_code_for_paper("2110.02037")
get_code_for_paper("2110.02037")
have("2110.02037")

The run record, dated, one paper per request, free:

curl https://syntology.ai/api/ran/2110.02037.json

A badge for a README (the split and the date, never a ratio):

[![Syntology run record](https://syntology.ai/api/ran/2110.02037.svg)](https://syntology.ai/paper/2110.02037)

Connect an agent — have() is free.