SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2412.09764 · 2024

Memory Layers at Scale

arXiv · PDF · Open in the Atlas

Code that ran

14 of 19 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; 5 honoured the contract we drafted, 0 violated it and 0 have no contract verdict; 9 failures of our instrument; 5 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/2412.09764.json · badge: /api/ran/2412.09764.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
facebookresearch/memory canonical (classifier label)
classifier basis: census
14 ran · 5 no recorded run
FunctionStatusWhere it lives
allclose Ran
ran; our contract draft was wrong, not the code
facebookresearch/memory/validate_distributed_embeddingbag.py
pointer only (licence: NOASSERTION) · get_code("3eb31c04dd0c9897")
How it ran

The body of this sample is withheld: GitHub could not classify the licence of this paper's repository, facebookresearch/memory (NOASSERTION). 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.

dataclass_from_dict Ran
ran; our contract draft was wrong, not the code
facebookresearch/memory/lingua/args.py
pointer only (licence: NONE) · get_code("e0ebc4027f1bd03b")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, facebookresearch/memory 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.

dataclass_to_dict Ran
ran; our contract draft was wrong, not the code
facebookresearch/memory/lingua/args.py
pointer only (licence: NONE) · get_code("116b2f19f9a6148c")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, facebookresearch/memory 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.

distribute_data_to_rank Ran
ran; our contract draft was wrong, not the code
facebookresearch/memory/lingua/data.py
pointer only (licence: NOASSERTION) · get_code("faca6fc3f7126a93")
How it ran

The body of this sample is withheld: GitHub could not classify the licence of this paper's repository, facebookresearch/memory (NOASSERTION). 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.

find_and_sanitize_chunks Ran
ran; our contract draft was wrong, not the code
facebookresearch/memory/lingua/data.py
pointer only (licence: NOASSERTION) · get_code("3c8fff70712ff33a")
How it ran

The body of this sample is withheld: GitHub could not classify the licence of this paper's repository, facebookresearch/memory (NOASSERTION). 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.

flatten_dict Ran
ran; our contract draft was wrong, not the code
facebookresearch/memory/lingua/args.py
pointer only (licence: NONE) · get_code("15be2e56e073a228")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, facebookresearch/memory 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.

get_empty_buffer_state Ran
ran; our contract draft was wrong, not the code
facebookresearch/memory/lingua/data.py
pointer only (licence: NONE) · get_code("bfdcca22e67a0a67")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, facebookresearch/memory 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.

get_num_params Ran
ran, honoured the contract we drafted
facebookresearch/memory/lingua/metrics.py
pointer only (licence: NONE); the stored copy get_code serves, from facebookresearch/lst, is pointer only (licence: NOASSERTION) · get_code("81d8582d1fab7bda")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, facebookresearch/memory 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_stats Ran
ran, honoured the contract we drafted
facebookresearch/memory/lingua/probe.py
pointer only (licence: NONE) · get_code("22fbaa2caf605450")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, facebookresearch/memory 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.

lr_cosine Ran
ran, honoured the contract we drafted
facebookresearch/memory/lingua/optim.py
pointer only (licence: NOASSERTION) · get_code("758fe7262f8fa41e")
How it ran

The body of this sample is withheld: GitHub could not classify the licence of this paper's repository, facebookresearch/memory (NOASSERTION). 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.

lr_inv_sqrt Ran
ran, honoured the contract we drafted
facebookresearch/memory/lingua/optim.py
pointer only (licence: NONE) · get_code("90403d20388a5636")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, facebookresearch/memory 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.

lr_linear Ran
ran, honoured the contract we drafted
facebookresearch/memory/lingua/optim.py
pointer only (licence: NONE) · get_code("4e2454a79a8338d0")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, facebookresearch/memory 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.

named_replace Ran
ran; our contract draft was wrong, not the code
facebookresearch/memory/lingua/float8.py
pointer only (licence: NONE); the stored copy get_code serves, from facebookresearch/lst, is pointer only (licence: NOASSERTION) · get_code("aaa4b37e564944fe")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, facebookresearch/memory 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.

scale Ran
ran; our contract draft was wrong, not the code
facebookresearch/memory/lingua/float8.py
pointer only (licence: NONE); the stored copy get_code serves, from facebookresearch/lst, is pointer only (licence: NOASSERTION) · get_code("2848341791f9a450")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, facebookresearch/memory 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.

consolidate_checkpoints Not yet run
unverified (harvested, no recorded run)
facebookresearch/memory/lingua/checkpoint.py
pointer only (licence: NOASSERTION) · get_code("be7ab0634167f84c")
How it ran

The body of this sample is withheld: GitHub could not classify the licence of this paper's repository, facebookresearch/memory (NOASSERTION). 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.

dist_max Not yet run
unverified (harvested, no recorded run)
facebookresearch/memory/lingua/distributed.py
pointer only (licence: NONE) · get_code("9caf44ff8e00b1dc")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, facebookresearch/memory 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. No invocation recorded.

dist_mean Not yet run
unverified (harvested, no recorded run)
facebookresearch/memory/lingua/distributed.py
pointer only (licence: NONE) · get_code("d09c7072d82024f4")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, facebookresearch/memory 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. No invocation recorded.

get_device_mesh Not yet run
unverified (harvested, no recorded run)
facebookresearch/memory/lingua/distributed.py
pointer only (licence: NOASSERTION) · get_code("eae71867f83f5695")
How it ran

The body of this sample is withheld: GitHub could not classify the licence of this paper's repository, facebookresearch/memory (NOASSERTION). 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.

matmul Not yet run
unverified (harvested, no recorded run)
facebookresearch/memory/lingua/float8.py
pointer only (licence: NONE); the stored copy get_code serves, from facebookresearch/lst, is pointer only (licence: NOASSERTION) · get_code("30e68c9cadfddd32")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, facebookresearch/memory 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. No invocation recorded.

Also present in other papers' code

Of 19 samples lifted from this paper's linked repositories, 19 carry a function identity and 0 do not. Of the 19 distinct functions with an identity, 14 are also present in other papers whose code we hold (1 with at least one paper dated earlier, 5 the same month, 14 later) and 5 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
flatten_dict 5 other papers
2 dated earlier, 3 later
2405.15589 (2024-05)
Efficient Adversarial Training in LLMs with Continuous Attacks
2410.15221 (2024-10)
IntersectionZoo: Eco-driving for Benchmarking Multi-Agent Contextual Reinforcement Learning
2503.18866 (2025-03)
Reasoning to Learn from Latent Thoughts
and 2 more not named here
dist_max 4 other papers
1 the same month, 3 later
2412.09871 (2024-12)
Byte Latent Transformer: Patches Scale Better Than Tokens
2503.18866 (2025-03)
Reasoning to Learn from Latent Thoughts
2506.14761 (2025-06)
From Bytes to Ideas: Language Modeling with Autoregressive U-Nets
and 1 more not named here
get_device_mesh 4 other papers
1 the same month, 3 later
2412.09871 (2024-12)
Byte Latent Transformer: Patches Scale Better Than Tokens
2503.18866 (2025-03)
Reasoning to Learn from Latent Thoughts
2506.14761 (2025-06)
From Bytes to Ideas: Language Modeling with Autoregressive U-Nets
and 1 more not named here
dataclass_from_dict 3 other papers
3 later
2503.18866 (2025-03)
Reasoning to Learn from Latent Thoughts
2506.14761 (2025-06)
From Bytes to Ideas: Language Modeling with Autoregressive U-Nets
2605.20670 (2026-05)
LT2: Linear-Time Looped Transformers
dataclass_to_dict 3 other papers
3 later
2503.18866 (2025-03)
Reasoning to Learn from Latent Thoughts
2506.14761 (2025-06)
From Bytes to Ideas: Language Modeling with Autoregressive U-Nets
2605.20670 (2026-05)
LT2: Linear-Time Looped Transformers
get_empty_buffer_state 3 other papers
3 later
2503.18866 (2025-03)
Reasoning to Learn from Latent Thoughts
2506.14761 (2025-06)
From Bytes to Ideas: Language Modeling with Autoregressive U-Nets
2605.20670 (2026-05)
LT2: Linear-Time Looped Transformers
get_num_params 3 other papers
3 later
2503.18866 (2025-03)
Reasoning to Learn from Latent Thoughts
2506.14761 (2025-06)
From Bytes to Ideas: Language Modeling with Autoregressive U-Nets
2510.06195 (2025-10)
Latent Speech-Text Transformer
matmul 3 other papers
1 the same month, 2 later
2412.09871 (2024-12)
Byte Latent Transformer: Patches Scale Better Than Tokens
2503.18866 (2025-03)
Reasoning to Learn from Latent Thoughts
2510.06195 (2025-10)
Latent Speech-Text Transformer
named_replace 3 other papers
1 the same month, 2 later
2412.09871 (2024-12)
Byte Latent Transformer: Patches Scale Better Than Tokens
2503.18866 (2025-03)
Reasoning to Learn from Latent Thoughts
2510.06195 (2025-10)
Latent Speech-Text Transformer
scale 3 other papers
1 the same month, 2 later
2412.09871 (2024-12)
Byte Latent Transformer: Patches Scale Better Than Tokens
2503.18866 (2025-03)
Reasoning to Learn from Latent Thoughts
2510.06195 (2025-10)
Latent Speech-Text Transformer
dist_mean 2 other papers
2 later
2503.18866 (2025-03)
Reasoning to Learn from Latent Thoughts
2506.14761 (2025-06)
From Bytes to Ideas: Language Modeling with Autoregressive U-Nets
log_stats 2 other papers
2 later
2503.18866 (2025-03)
Reasoning to Learn from Latent Thoughts
2506.14761 (2025-06)
From Bytes to Ideas: Language Modeling with Autoregressive U-Nets
lr_inv_sqrt 2 other papers
2 later
2503.18866 (2025-03)
Reasoning to Learn from Latent Thoughts
2506.14761 (2025-06)
From Bytes to Ideas: Language Modeling with Autoregressive U-Nets
lr_linear 2 other papers
2 later
2503.18866 (2025-03)
Reasoning to Learn from Latent Thoughts
2506.14761 (2025-06)
From Bytes to Ideas: Language Modeling with Autoregressive U-Nets

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

Memory layers use a trainable key-value lookup mechanism to add extra parameters to a model without increasing FLOPs. Conceptually, sparsely activated memory layers complement compute-heavy dense feed-forward layers, providing dedicated capacity to store and retrieve information cheaply. This work takes memory layers beyond proof-of-concept, proving their utility at contemporary scale. On downstream tasks, language models augmented with our improved memory layer outperform dense models with more than twice the computation budget, as well as mixture-of-expert models when matched for both compute and parameters. We find gains are especially pronounced for factual tasks. We provide a fully parallelizable memory layer implementation, demonstrating scaling laws with up to 128B memory parameters, pretrained to 1 trillion tokens, comparing to base models with up to 8B parameters.

For agents

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

get_harvested_code_for_paper("2412.09764")
get_code_for_paper("2412.09764")
have("2412.09764")

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

curl https://syntology.ai/api/ran/2412.09764.json

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

[![Syntology run record](https://syntology.ai/api/ran/2412.09764.svg)](https://syntology.ai/paper/2412.09764)

Connect an agent — have() is free.