SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2502.01591 · ICML · 2025

Improving Transformer World Models for Data-Efficient RL

Joseph Ortiz, Kevin Murphy, Antoine Dedieu, Wolfgang Lehrach, Carter Wendelken, Xinghua Lou, J Swaroop Guntupalli, Miguel Lazaro-Gredilla

arXiv · PDF · Open in the Atlas

Code that ran

9 of 13 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (0 of them constructed an object rather than computing a result; 0 honoured the contract we drafted, 0 violated it and 2 have no contract verdict; 7 failures of our instrument; 4 no recorded run), measured 2026-11-22; 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/2502.01591.json · badge: /api/ran/2502.01591.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
luchris429/purejaxrl canonical (classifier label)
classifier basis not recorded on the link
3 ran · 0 no recorded run
eloialonso/iris community
Papers with Code archive
6 ran · 4 no recorded run
FunctionStatusWhere it lives
ScalingLayer Ran
ran on a synthesized input
eloialonso/iris/src/models/tokenizer/tokenizer.py
pointer only (licence: GPL-3.0) · get_code("878ced2dad28f2bb")
How it ran

The body of this sample is withheld: GPL-3.0, as GitHub detects it on this paper's repository, eloialonso/iris, is not in the permissive set Syntology serves inline. 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.

get_ckpt_path Ran
ran; our contract draft was wrong, not the code
eloialonso/iris/src/models/tokenizer/tokenizer.py
pointer only (licence: GPL-3.0) · get_code("970bb08b535c4c9e")
How it ran

The body of this sample is withheld: GPL-3.0, as GitHub detects it on this paper's repository, eloialonso/iris, is not in the permissive set Syntology serves inline. 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.

init_VinvB Ran
ran; our contract draft was wrong, not the code
luchris429/purejaxrl/purejaxrl/experimental/s5/s5.py
pointer only (licence: Apache-2.0); the stored copy get_code serves, from luchris429/popjaxrl, is pointer only (licence: NONE) · get_code("ec8a321400aae8d2")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (Apache-2.0), but get_code serves the stored copy, first harvested from luchris429/popjaxrl, and GitHub detected no licence in that repository 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.

init_log_steps Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
luchris429/purejaxrl/purejaxrl/experimental/s5/s5.py
pointer only (licence: Apache-2.0); the stored copy get_code serves, from luchris429/popjaxrl, is pointer only (licence: NONE) · get_code("a4bdc24302b8ab98")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (Apache-2.0), but get_code serves the stored copy, first harvested from luchris429/popjaxrl, and GitHub detected no licence in that repository 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.

log_step_initializer Ran
ran; our contract draft was wrong, not the code
luchris429/purejaxrl/purejaxrl/experimental/s5/s5.py
pointer only (licence: Apache-2.0); the stored copy get_code serves, from luchris429/popjaxrl, is pointer only (licence: NONE) · get_code("8b1a3bc2f6b64dce")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (Apache-2.0), but get_code serves the stored copy, first harvested from luchris429/popjaxrl, and GitHub detected no licence in that repository 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.

md5_hash Ran
ran; our contract draft was wrong, not the code
eloialonso/iris/src/models/tokenizer/tokenizer.py
pointer only (licence: GPL-3.0) · get_code("ebaf93caaa927ff9")
How it ran

The body of this sample is withheld: GPL-3.0, as GitHub detects it on this paper's repository, eloialonso/iris, is not in the permissive set Syntology serves inline. 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.

normalize_tensor Ran
ran; our contract draft was wrong, not the code
eloialonso/iris/src/models/tokenizer/tokenizer.py
pointer only (licence: GPL-3.0) · get_code("4d9196e5759e6950")
How it ran

The body of this sample is withheld: GPL-3.0, as GitHub detects it on this paper's repository, eloialonso/iris, is not in the permissive set Syntology serves inline. 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.

spatial_average Ran
ran; our contract draft was wrong, not the code
eloialonso/iris/src/models/tokenizer/tokenizer.py
pointer only (licence: GPL-3.0) · get_code("6855e475947477db")
How it ran

The body of this sample is withheld: GPL-3.0, as GitHub detects it on this paper's repository, eloialonso/iris, is not in the permissive set Syntology serves inline. 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.

vgg16 Ran
ran on a synthesized input
eloialonso/iris/src/models/tokenizer/tokenizer.py
pointer only (licence: GPL-3.0) · get_code("ca3575bced071dc5")
How it ran

The body of this sample is withheld: GPL-3.0, as GitHub detects it on this paper's repository, eloialonso/iris, is not in the permissive set Syntology serves inline. 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.

LPIPS Not yet run
unverified (harvested, no recorded run)
eloialonso/iris/src/models/tokenizer/tokenizer.py
pointer only (licence: GPL-3.0) · get_code("4767c4ed7a66b542")
How it ran

The body of this sample is withheld: GPL-3.0, as GitHub detects it on this paper's repository, eloialonso/iris, is not in the permissive set Syntology serves inline. 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. No invocation recorded.

NetLinLayer Not yet run
unverified (harvested, no recorded run)
eloialonso/iris/src/models/tokenizer/tokenizer.py
pointer only (licence: GPL-3.0) · get_code("8681e7e0f15169da")
How it ran

The body of this sample is withheld: GPL-3.0, as GitHub detects it on this paper's repository, eloialonso/iris, is not in the permissive set Syntology serves inline. 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. No invocation recorded.

Tokenizer Not yet run
unverified (harvested, no recorded run)
eloialonso/iris/src/models/tokenizer/tokenizer.py
pointer only (licence: GPL-3.0) · get_code("a823cb01993404cf")
How it ran

The body of this sample is withheld: GPL-3.0, as GitHub detects it on this paper's repository, eloialonso/iris, is not in the permissive set Syntology serves inline. 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. No invocation recorded.

download Not yet run
unverified (harvested, no recorded run)
eloialonso/iris/src/models/tokenizer/tokenizer.py
pointer only (licence: GPL-3.0) · get_code("5b78fedd249dc684")
How it ran

The body of this sample is withheld: GPL-3.0, as GitHub detects it on this paper's repository, eloialonso/iris, is not in the permissive set Syntology serves inline. 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. No invocation recorded.

Also present in other papers' code

Of 13 samples lifted from this paper's linked repositories, 13 carry a function identity and 0 do not. Of the 13 distinct functions with an identity, 5 are also present in other papers whose code we hold (5 with at least one paper dated earlier, 0 the same month, 0 later) and 8 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.

FunctionAlso present inEarliest-datedOthers named
Tokenizer
brick of this paper's assembly: entry
1 other paper
1 dated earlier
2209.00588 (2022-09)
Transformers are Sample Efficient World Models
init_VinvB 1 other paper
1 dated earlier
2303.03982 (2023-03)
Structured State Space Models for In-Context Reinforcement Learning
init_log_steps 1 other paper
1 dated earlier
2303.03982 (2023-03)
Structured State Space Models for In-Context Reinforcement Learning
log_step_initializer 1 other paper
1 dated earlier
2303.03982 (2023-03)
Structured State Space Models for In-Context Reinforcement Learning
normalize_tensor
brick of this paper's assembly: helper
1 other paper
1 dated earlier
2307.15208 (2023-07)
Generative AI for Medical Imaging: extending the MONAI Framework

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 present three improvements to the standard model-based RL paradigm based on transformers: (a) "Dyna with warmup", which trains the policy on real and imaginary data, but only starts using imaginary data after the world model has been sufficiently trained; (b) "nearest neighbor tokenizer" for image patches, which improves upon previous tokenization schemes, which are needed when using a transformer world model (TWM), by ensuring the code words are static after creation, thus providing a constant target for TWM learning; and (c) "block teacher forcing", which allows the TWM to reason jointly about the future tokens of the next timestep, instead of generating them sequentially. We then show that our method significantly improves upon prior methods in various environments. We mostly focus on the challenging Craftax-classic benchmark, where our method achieves a reward of 69.66% after only 1M environment steps, significantly outperforming DreamerV3, which achieves 53.2%, and exceeding human performance of 65.0% for the first time. We also show preliminary results on Craftax-full, MinAtar, and three different two-player games, to illustrate the generality of the approach.

For agents

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

get_harvested_code_for_paper("2502.01591")
get_code_for_paper("2502.01591")
have("2502.01591")

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

curl https://syntology.ai/api/ran/2502.01591.json

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

[![Syntology run record](https://syntology.ai/api/ran/2502.01591.svg)](https://syntology.ai/paper/2502.01591)

Connect an agent — have() is free.