SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2601.17883 · 2026

EEG-FM-Compass: Progress, Benchmarking, and Future Directions for EEG Foundation Models

Dongrui Wu, Encoder Decoder, Yuheng Chen, Yaozhi Wen, Zhenyao Cui, Dingkun Liu, Zhu Chen, Jingwei Luo, Jiayu An

arXiv · PDF · Open in the Atlas

Code that ran

10 of 11 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 4 have no contract verdict; 6 failures of our instrument; 1 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/2601.17883.json · badge: /api/ran/2601.17883.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
Dingkun0817/EEG-FM-Benchmark community
Papers with Code archive
10 ran · 1 no recorded run
FunctionStatusWhere it lives
Attention Ran
ran on a synthesized input · constructed an object; no result was computed
Dingkun0817/EEG-FM-Benchmark/models/FM/EEGPT/Model_EEGPT.py
code served (permissive licence) · get_code("fb2c57145dd97d77")
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, Attention: 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
import math

torch.manual_seed(0)

# Define a small frequency tensor for rotary embeddings
# Shape: (seq_len, head_dim) where head_dim must be even for rotate_half to work
freqs = torch.randn(3, 2)  # seq_len=3, head_dim=2

# Define input tensor
# Shape: (batch, seq_len, dim)
x = torch.randn(2, 3, 4)  # batch=2, seq_len=3, dim=4

ARGS = {
    "dim": 4,
    "num_heads": 2,
    "qkv_bias": False,
    "attn_drop": 0.0,
    "proj_drop": 0.0,
    "is_causal": False,
    "use_rope": False,
    "return_attention": 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 Attention — 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 Attention 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
Block Ran
ran on a synthesized input
Dingkun0817/EEG-FM-Benchmark/models/FM/EEGPT/Model_EEGPT.py
code served (permissive licence) · get_code("6399e309ecb84ffc")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, class construct-then-call prompt, model not recorded) and returned Tensor, ndim 3, dtype family float; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim 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 math

# Helper function for rotary embeddings (copied from the class definition)
def apply_rotary_emb(freqs, t, start_index=0, scale=1.0):
    freqs = freqs.to(t.device)
    rot_dim = freqs.shape[-1]
    end_index = start_index + rot_dim
    assert rot_dim <= t.shape[-1], f'feature dimension {t.shape[-1]} is not of sufficient size to rotate in all the positions {rot_dim}'
    t_left, t_middle, t_right = (t[..., :start_index], t[..., start_index:end_index], t[..., end_index:])
    t_rotated_middle = t_middle * freqs.cos() * scale + rotate_half(t_middle) * freqs.sin() * scale
    return torch.cat((t_left, t_rotated_middle, t_right), dim=-1)

def rotate_half(x):
    x = x.reshape((*x.shape[:-1], x.shape[-1] // 2, 2))
    x1, x2 = x.unbind(dim=-1)
    x = torch.stack((-x2, x1), dim=-1)
    return x.flatten(-2)

# Define INIT_ARGS for Block
INIT_ARGS = {
    'dim': 4,
    'num_heads': 2,
    'mlp_ratio': 2.0,
    'qkv_bias': False,
    'drop': 0.0,
    'attn_drop': 0.0,
    'drop_path': 0.0,
    'act_layer': nn.GELU,
    'norm_layer': nn.LayerNorm,
    'is_causal': False,
    'use_rope': False,
    'return_attention': False
}

# Create a small input tensor and rotary frequencies (if needed)
B, T, C = 2, 3, 4  # batch=2, sequence=3, features=4
x = torch.randn(B, T, C)

# For rotary embeddings, create dummy frequencies if use_rope is True
freqs = torch.randn(T, C // 2) if INIT_ARGS['use_rope'] else None

# Define ARGS for Block.forward
ARGS = {
    'x': x,
    'freqs': freqs
}

# Patch the missing norm_layer in Block.__init__ (norm_la -> norm_layer)
class Block(nn.Module):
    def __init__(self, dim, num_heads, mlp_ratio=4.0, qkv_bias=False, drop=0.0, attn_drop=0.0, drop_path=0.0, act_layer=nn.GELU, norm_layer=nn.LayerNorm, is_causal=False, use_rope=False, return_attention=False):
        super().__init__()
        self.return_attention = return_attention
        self.norm1 = norm_layer(dim)
        self.attn = Attention(dim, num_heads=num_heads, qkv_bias=qkv_bias, attn_drop=attn_drop, proj_drop=drop, is_causal=is_causal, use_rope=use_rope, return_attention=return_attention)
        self.drop_path = DropPath(drop_path) if drop_path > 0.0 else nn.Identity()
        self.norm2 = norm_layer(dim)
        mlp_hidden_dim = int(dim * mlp_ratio)
        self.mlp = MLP(in_features=dim, hidden_features=mlp_hidden_dim, act_layer=act_layer, drop=drop)

    def forward(self, x, freqs=None):
        if self.return_attention:
            attn = self.attn(self.norm1(x), freqs=freqs)
            return attn
        x = x + self.drop_path(self.attn(self.norm1(x), freqs=freqs))
        x = x + self.drop_path(self.mlp(self.norm2(x)))
        return x
Setup block drafted by
a language model, class construct-then-call prompt, model not recorded
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Tensor, ndim 3, dtype family float — no output shape is stated: observe_signature.py observed this class with its constructor sizes canonicalised where it could, which the node does not record, so its shape is not stated as this setup block's
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
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: torch and numpy seeded to 0 by the harness immediately before the class was constructed from INIT_ARGS (fixture_check.py)
DropPath Ran
ran on a synthesized input · constructed an object; no result was computed
Dingkun0817/EEG-FM-Benchmark/models/FM/EEGPT/Model_EEGPT.py
code served (permissive licence) · get_code("8e464eea7a15e985")
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, DropPath: 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)

# DropPath only uses:
# - __init__ parameter: drop_prob (float or None)
# - forward method which calls drop_path with x (a tensor)
# The drop_path method uses:
#   - x.shape, x.ndim, x.dtype, x.device
#   - self.training attribute (set by nn.Module)
#   - tensor operations: torch.rand, floor_, div

# Create a small input tensor: batch_size=2, feature_dim=4
x = torch.randn(2, 4)

# Create DropPath instance with drop_prob
drop_prob_value = 0.2

ARGS = {
    "drop_prob": drop_prob_value,
}
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 DropPath — 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 DropPath 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
PatchEmbed Ran
ran on a synthesized input · constructed an object; no result was computed
Dingkun0817/EEG-FM-Benchmark/models/FM/EEGPT/Model_EEGPT.py
code served (permissive licence) · get_code("cc322c6117069ffd")
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, PatchEmbed: 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)

ARGS = {
    "img_size": (64, 1000),
    "patch_size": 16,
    "patch_stride": None,
    "embed_dim": 768,
}
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 PatchEmbed — 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 PatchEmbed 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
_no_grad_trunc_normal_ Ran
ran; our contract draft was wrong, not the code
Dingkun0817/EEG-FM-Benchmark/models/FM/EEGPT/Model_EEGPT.py
code served (permissive licence) · get_code("307242ac3279e182")
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 Tensor, ndim 2, dtype family float, shape (2, 4) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape 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 math

torch.manual_seed(0)

# Create a small tensor to be initialized
tensor = torch.empty(2, 4)

# Define parameters for truncated normal distribution
mean = 0.0
std = 1.0
a = -2.0  # lower bound
b = 2.0   # upper bound

ARGS = {
    "tensor": tensor,
    "mean": mean,
    "std": std,
    "a": a,
    "b": b,
}
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 Tensor, ndim 2, dtype family float, shape (2, 4) — shape (2, 4) as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
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
apply_mask Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
Dingkun0817/EEG-FM-Benchmark/models/FM/EEGPT/Model_EEGPT.py
code served (permissive licence) · get_code("4265c294be96833d")
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 Tensor, ndim 4, dtype family float, shape (2, 2, 3, 4) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape 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

torch.manual_seed(0)

# Case 1: 2D mask (the main case in the function)
# x shape: [B, N, C, D] = [2, 5, 3, 4]
# mask shape: [mN, mC] where mN*mC indices fit within N*C
B, N, C, D = 2, 5, 3, 4
mN, mC = 3, 2

x = torch.randn(B, N, C, D)

# Create a 2D mask with valid indices into the flattened N*C dimension
# Indices should be < N*C = 15
mask = torch.tensor([
    [0, 2, 5],
    [7, 10, 14]
], dtype=torch.long)

ARGS = {
    "mask": mask,
    "x": x
}
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 Tensor, ndim 4, dtype family float, shape (2, 2, 3, 4) — shape (2, 2, 3, 4) as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
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
apply_mask_t Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
Dingkun0817/EEG-FM-Benchmark/models/FM/EEGPT/Model_EEGPT.py
code served (permissive licence) · get_code("3a7b9787a11d5c2f")
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 Tensor, ndim 3, dtype family float, shape (2, 3, 4) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape 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

torch.manual_seed(0)

# Construct arguments for apply_mask_t
# x: tensor of shape [B, N, D] where B=batch-size, N=num-patches, D=feature-dim
B = 2  # batch size
N = 5  # num patches
D = 4  # feature dimension

x = torch.randn(B, N, D)

# mask_t: tensor of shape [mN, mC] containing indices
# Looking at the code: mask_t.reshape((1, mN, 1)) suggests mN is the first dimension
# The gather operation uses mask_keep with shape [B, mN, D]
# So mask_t should have shape [mN] or [mN, mC] where mC=1
mN = 3  # number of patches to keep
mask_t = torch.tensor([0, 2, 4])  # indices into the N dimension

ARGS = {
    "mask_t": mask_t,
    "x": x,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Tensor, ndim 3, dtype family float, shape (2, 3, 4) — shape (2, 3, 4) as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
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
apply_rotary_emb Ran
ran; our contract draft was wrong, not the code
Dingkun0817/EEG-FM-Benchmark/models/FM/EEGPT/Model_EEGPT.py
code served (permissive licence) · get_code("db94a8e325ace4ed")
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 Tensor, ndim 3, dtype family float, shape (2, 3, 6) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape 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)

# Create frequency embeddings (precomputed sine and cosine values)
# Shape: (rot_dim,) where rot_dim is the rotation dimension
freqs = torch.randn(4)  # 4-dimensional rotation frequencies

# Create input tensor with sufficient size
# Shape: (batch_size, seq_len, feature_dim)
# We need feature_dim >= rot_dim (4) for the assertion to pass
t = torch.randn(2, 3, 6)  # batch=2, seq_len=3, feature_dim=6

# start_index determines where rotation begins
start_index = 0

# scale factor for rotation
scale = 1.0

ARGS = {
    "freqs": freqs,
    "t": t,
    "start_index": start_index,
    "scale": scale,
}
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 Tensor, ndim 3, dtype family float, shape (2, 3, 6) — shape (2, 3, 6) as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
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
rotate_half Ran
ran; our contract draft was wrong, not the code
Dingkun0817/EEG-FM-Benchmark/models/FM/EEGPT/Model_EEGPT.py
code served (permissive licence) · get_code("c1f873b79cb6cd24")
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 Tensor, ndim 2, dtype family float, shape (2, 4) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape 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

torch.manual_seed(0)

# The function rotate_half takes a tensor x and:
# 1. Reshapes it to (..., x.shape[-1] // 2, 2)
# 2. Unbinds along the last dimension
# 3. Stacks them back in rotated order
# 4. Flattens back to original shape

# So x needs to have an even last dimension
# Let's create a small tensor with batch size 2 and last dimension 4

x = torch.randn(2, 4)

ARGS = {
    "x": x
}
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 Tensor, ndim 2, dtype family float, shape (2, 4) — shape (2, 4) as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
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
trunc_normal_ Ran
ran; our contract draft was wrong, not the code
Dingkun0817/EEG-FM-Benchmark/models/FM/EEGPT/Model_EEGPT.py
code served (permissive licence) · get_code("359975e35f7d1d79")
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 Tensor, ndim 2, dtype family float, shape (2, 4) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape 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 math

torch.manual_seed(0)

# Create a small tensor to be initialized
tensor = torch.empty(2, 4)

# Define arguments for trunc_normal_
ARGS = {
    "tensor": tensor,
    "mean": 0.0,
    "std": 1.0,
    "a": -2.0,
    "b": 2.0
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Tensor, ndim 2, dtype family float, shape (2, 4) — shape (2, 4) as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
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
EEGTransformer Not yet run
unverified (harvested, no recorded run)
Dingkun0817/EEG-FM-Benchmark/models/FM/EEGPT/Model_EEGPT.py
code served (permissive licence) · get_code("a61aafbddb64d2d8")
How it ran

no invocation recorded.

Also present in other papers' code

Of 11 samples lifted from this paper's linked repositories, 11 carry a function identity and 0 do not. Of the 11 distinct functions with an identity, 4 are also present in other papers whose code we hold (4 with at least one paper dated earlier, 1 the same month, 3 later) and 7 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
trunc_normal_
brick of this paper's assembly: helper
160 other papers
150 dated earlier, 3 the same month, 7 later · 36 only as an assembly brick
1911.10194 (2019-11)
Panoptic-DeepLab: A Simple, Strong, and Fast Baseline for Bottom-Up Panoptic Segmentation
2005.10242 (2020-05)
Understanding Contrastive Representation Learning through Alignment and Uniformity on the Hypersphere
2007.12911 (2020-07)
Tighter risk certificates for neural networks
and 157 more not named here
apply_mask
brick of this paper's assembly: helper
2 other papers
1 dated earlier, 1 later
2508.17742 (2025-08)
EEG-FM-Bench: A Comprehensive Benchmark for the Systematic Evaluation and Diagnostic Analyses of EEG Foundation Models
2609.15687 (2026-09)
EEG-Xplain: Decoding Neural Black-Boxes of EEG Foundation Models
apply_mask_t
brick of this paper's assembly: helper
2 other papers
1 dated earlier, 1 later
2508.17742 (2025-08)
EEG-FM-Bench: A Comprehensive Benchmark for the Systematic Evaluation and Diagnostic Analyses of EEG Foundation Models
2609.15687 (2026-09)
EEG-Xplain: Decoding Neural Black-Boxes of EEG Foundation Models
_no_grad_trunc_normal_
brick of this paper's assembly: helper
1 other paper
1 dated earlier · 1 only as an assembly brick
2301.08243 (2023-01)
Self-Supervised Learning from Images with a Joint-Embedding Predictive Architecture

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

Electroencephalography (EEG) foundation models (FMs) have recently emerged as a promising paradigm for brain-computer interfaces, aiming to learn transferable neural representations from large-scale heterogeneous recordings. Despite rapid progress, a fair and comprehensive comparison of existing EEG FMs is still lacking, owing to inconsistent pre-training objectives, preprocessing choices, and downstream evaluation protocols. To fill this gap, we present EEG-FM-Compass. We first review 55 representative models and organize their design choices into a unified taxonomic framework including data standardization, model architectures, and self-supervised pre-training strategies. We then evaluate 12 open source FMs and competitive specialist baselines across 13 EEG datasets spanning nine brain-computer interface paradigms. Emphasizing real-world deployments, we consider both cross-subject generalization under a leave-one-subject-out protocol and rapid calibration under a within-subject few-shot setting. We further compare full-parameter fine-tuning with linear probing to assess the transferability of pre-trained representations, and examine the relationship between model scale and downstream performance. Our results indicate that: 1) linear probing is frequently insufficient; 2) specialist models trained from scratch remain competitive across many tasks; and 3) larger FMs do not necessarily yield better generalization performance under current data regimes and training practices.

For agents

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

get_harvested_code_for_paper("2601.17883")
get_code_for_paper("2601.17883")
have("2601.17883")

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

curl https://syntology.ai/api/ran/2601.17883.json

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

[![Syntology run record](https://syntology.ai/api/ran/2601.17883.svg)](https://syntology.ai/paper/2601.17883)

Connect an agent — have() is free.