SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2605.04651 · 2026

FAAST: Forward-Only Associative Learning via Closed-Form Fast Weights for Test-Time Supervised Adaptation

Ke Sun, Yue Zhang, Han Cui, Guangsheng Bao, Yanbin Zhao, Hongbo Zhang, Juncai He

arXiv · PDF

Code that ran

3 of 6 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (3 of them constructed an object rather than computing a result; 0 honoured the contract we drafted, 0 violated it and 3 have no contract verdict; 0 failures of our instrument; 3 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/2605.04651.json · badge: /api/ran/2605.04651.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
baoguangsheng/faast community
Papers with Code archive
3 ran · 3 no recorded run
FunctionStatusWhere it lives
InverseReader Ran
ran on a synthesized input · constructed an object; no result was computed
baoguangsheng/faast/llm/memory/memory_layer.py
code served (permissive licence) · get_code("886d738af359fde6")
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, through the batch API) and returned an instance of its own class, InverseReader: the run constructed the object and did not call it, so no result was computed. The image it ran in was 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

torch.manual_seed(0)

class MockConfig:
    def __init__(self):
        self.hidden_size = 4
        self.reader_update_discount = 0.9
        self.reader_update_alpha = 0.5
        self.reader_update_size = 1

ARGS = {
    "config": MockConfig(),
    "layer_index": 0,
    "is_output_layer": False,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type InverseReader — the run constructed an instance of its own class and did not call it; no shape is stated — the instrument coerces any output to an array; for a InverseReader output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not 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
MemoryBank Ran
ran on a synthesized input · constructed an object; no result was computed
baoguangsheng/faast/llm/memory/memory_layer.py
code served (permissive licence) · get_code("b1e0da9a2ab1e77e")
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, through the batch API) and returned an instance of its own class, MemoryBank: the run constructed the object and did not call it, so no result was computed. The image it ran in was 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

torch.manual_seed(0)

class Config:
    """Minimal config class for MemoryBank."""
    def __init__(self):
        self.hidden_dim = 4
        self.num_layers = 2

config = Config()
layer_index = 0
is_output_layer = False

ARGS = {
    "config": config,
    "layer_index": layer_index,
    "is_output_layer": is_output_layer,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type MemoryBank — the run constructed an instance of its own class and did not call it; no shape is stated — the instrument coerces any output to an array; for a MemoryBank output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not 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
ProjectionInterpreter Ran
ran on a synthesized input · constructed an object; no result was computed
baoguangsheng/faast/llm/memory/memory_layer.py
code served (permissive licence) · get_code("da6caff5507cc05c")
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, through the batch API) and returned an instance of its own class, ProjectionInterpreter: the run constructed the object and did not call it, so no result was computed. The image it ran in was 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

torch.manual_seed(0)

# Stand-in config class with required attributes
class Config:
    def __init__(self):
        self.hidden_size = 4

config = Config()
layer_index = 0
is_output_layer = False

# Create tensors with distinct shapes for probing
batch_size = 2
hidden_dim = 4
hidden_states = torch.randn(batch_size, hidden_dim)
mem_states = torch.randn(batch_size, hidden_dim)
attention_mask = torch.ones(batch_size, dtype=torch.bool)

ARGS = {
    "config": config,
    "layer_index": layer_index,
    "is_output_layer": is_output_layer,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ProjectionInterpreter — the run constructed an instance of its own class and did not call it; no shape is stated — the instrument coerces any output to an array; for a ProjectionInterpreter output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not 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
MemoryLayer Not yet run
unverified (harvested, no recorded run)
baoguangsheng/faast/llm/memory/memory_layer.py
code served (permissive licence) · get_code("3e8d947c01d33c35")
How it ran

no invocation recorded.

NextTokenExtractor Not yet run
unverified (harvested, no recorded run)
baoguangsheng/faast/llm/memory/memory_layer.py
code served (permissive licence) · get_code("2fa567ce3392ba94")
How it ran

no invocation recorded.

pinv_lowrank Not yet run
unverified (harvested, no recorded run)
baoguangsheng/faast/llm/memory/memory_layer.py
code served (permissive licence) · get_code("30e2958c2da7a3f0")
How it ran

no invocation recorded.

Also present in other papers' code

Of 6 samples lifted from this paper's linked repositories, 6 carry a function identity and 0 do not. Of the 6 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 6 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

Adapting pretrained models typically involves a trade-off between the high training costs of backpropagation and the heavy inference overhead of memory-based or in-context learning. We propose FAAST, a forward-only associative adaptation method that analytically compiles labeled examples into fast weights in a single pass. By eliminating memory or context dependence, FAAST achieves constant-time inference and decouples task adaptation from pretrained representation. Across image classification and language modeling benchmarks, FAAST matches or exceeds backprop-based adaptation while reducing adaptation time by over 90% and is competitive to memory/context-based adaptation while saving memory usage by up to 95%. These results demonstrate FAAST as a highly efficient, scalable solution for supervised task adaptation, particularly for resource-constrained models. We release the code and models at https: //github.com/baoguangsheng/faast.

For agents

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

get_harvested_code_for_paper("2605.04651")
get_code_for_paper("2605.04651")
have("2605.04651")

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

curl https://syntology.ai/api/ran/2605.04651.json

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

[![Syntology run record](https://syntology.ai/api/ran/2605.04651.svg)](https://syntology.ai/paper/2605.04651)

Connect an agent — have() is free.