SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2406.01538 · 2024

What Are Large Language Models Mapping to in the Brain? A Case Against Over-Reliance on Brain Scores

arXiv · PDF · Open in the Atlas

Code that ran

6 of 6 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, 2 violated it and 0 have no contract verdict; 4 failures 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/2406.01538.json · badge: /api/ran/2406.01538.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
ebrahimfeghhi/beyond-brainscore canonical (classifier label)
classifier basis: census
6 ran · 0 no recorded run
FunctionStatusWhere it lives
compute_p_value Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
ebrahimfeghhi/beyond-brainscore/misc_code/stats_testing_funcs.py
code served (permissive licence) · get_code("15d17754fadaeed5")
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 ndarray, ndim 1, dtype family float, shape (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 numpy as np

# Set random seed for reproducibility
np.random.seed(0)

# Define parameters
num_voxels = 4
num_permutations = 3

# Create real_r2: shape (num_voxels,)
real_r2 = np.array([0.5, 0.6, 0.4, 0.7])

# Create permuted_r2: shape (num_permutations, num_voxels)
# Make sure some permuted values are greater than real values to get non-zero p-values
permuted_r2 = np.array([
    [0.3, 0.5, 0.6, 0.5],
    [0.4, 0.7, 0.3, 0.8],
    [0.6, 0.4, 0.5, 0.6]
])

# Define the arguments
ARGS = {
    "real_r2": real_r2,
    "permuted_r2": permuted_r2,
    "num_permutations": num_permutations,
    "FDR": False
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, scipy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ndarray, ndim 1, dtype family float, shape (4,) — shape (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
is_punctuation Ran
ran, violated the contract we drafted
ebrahimfeghhi/beyond-brainscore/generate_activations/interp_funcs.py
code served (permissive licence) · get_code("cdbdae1ce67da8d6")
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 bool, ndim 0, dtype family float, shape () 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

def is_punctuation(char):
    sentence_ending_punctuation = ['.', '!', '?', ',']
    return char in sentence_ending_punctuation

ARGS = {
    'char': '.'
}
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 bool, ndim 0, dtype family float, shape () — shape () 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
is_sentence_ending_punctuation Ran
ran, violated the contract we drafted
ebrahimfeghhi/beyond-brainscore/generate_activations/interp_funcs.py
code served (permissive licence) · get_code("4e3b19fa43f47e91")
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 bool, ndim 0, dtype family float, shape () 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

def is_sentence_ending_punctuation(char):
    sentence_ending_punctuation = ['.', '!', '?']
    return char in sentence_ending_punctuation

ARGS = {
    'char': '.'
}
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 bool, ndim 0, dtype family float, shape () — shape () 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
permute_matrix_in_blocks Ran
ran; our contract draft was wrong, not the code
ebrahimfeghhi/beyond-brainscore/misc_code/stats_testing_funcs.py
code served (permissive licence) · get_code("9318f410cafebb3c")
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 tuple. 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 numpy as np
from sklearn.metrics import mean_squared_error
import copy

# Set random seed for reproducibility
np.random.seed(0)

# Create small test data
num_samples = 8  # Small number divisible by block_size
num_outputs = 3  # Number of output targets

# y_pred: predictions from model (num_samples x num_outputs)
y_pred = np.random.randn(num_samples, num_outputs).astype(np.float32)

# y_test: ground truth labels (num_samples x num_outputs)
y_test = np.random.randn(num_samples, num_outputs).astype(np.float32)

# mse_intercept_only: MSE of intercept-only baseline model (num_outputs,)
# This is the MSE when predicting just the mean
intercept_pred = np.mean(y_test, axis=0, keepdims=True)
mse_intercept_only = mean_squared_error(
    np.repeat(intercept_pred, num_samples, axis=0), 
    y_test, 
    multioutput='raw_values'
)

# block_size: size of blocks for permutation (should divide num_samples)
block_size = 2

# num_permutations: number of permutations to perform
num_permutations = 10

ARGS = {
    'y_pred': y_pred,
    'y_test': y_test,
    'mse_intercept_only': mse_intercept_only,
    'block_size': block_size,
    'num_permutations': num_permutations
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, sklearn — 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
r2_difference Ran
ran; our contract draft was wrong, not the code
ebrahimfeghhi/beyond-brainscore/misc_code/stats_testing_funcs.py
code served (permissive licence) · get_code("d3ccc4988979917b")
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 tuple, ndim 2, 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 numpy as np
from sklearn.metrics import mean_squared_error
import copy
from scipy.stats import false_discovery_control

np.random.seed(0)

# Create small test data
num_samples = 10
num_voxels = 4

# Generate predictions from two models
model1_preds = np.random.randn(num_samples, num_voxels).astype(np.float32)
model2_preds = np.random.randn(num_samples, num_voxels).astype(np.float32)

# Generate ground truth
ytest = np.random.randn(num_samples, num_voxels).astype(np.float32)

# Compute MSE for intercept-only model (mean prediction)
mean_pred = np.mean(ytest, axis=0, keepdims=True)
mse_intercept_only = np.mean((ytest - mean_pred) ** 2, axis=0)

ARGS = {
    'model1_preds': model1_preds,
    'model2_preds': model2_preds,
    'ytest': ytest,
    'mse_intercept_only': mse_intercept_only,
    'num_permutations': 10,
    'FDR': True,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, scipy, sklearn — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type tuple, ndim 2, dtype family float — a tuple output has no array shape of its own; none is stated
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
separate_words_commas_periods Ran
ran; our contract draft was wrong, not the code
ebrahimfeghhi/beyond-brainscore/generate_activations/interp_funcs.py
code served (permissive licence) · get_code("d089027452af38fb")
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 list, ndim 1, 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.

ARGS = {
    "sentence": "Hello, world! How are you?"
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
none outside the standard library — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type list, ndim 1, dtype family float — a list output has no array shape of its own; none is stated
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

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

Given the remarkable capabilities of large language models (LLMs), there has been a growing interest in evaluating their similarity to the human brain. One approach towards quantifying this similarity is by measuring how well a model predicts neural signals, also called "brain score". Internal representations from LLMs achieve state-of-the-art brain scores, leading to speculation that they share computational principles with human language processing. This inference is only valid if the subset of neural activity predicted by LLMs reflects core elements of language processing. Here, we question this assumption by analyzing three neural datasets used in an impactful study on LLM-to-brain mappings, with a particular focus on an fMRI dataset where participants read short passages. We first find that when using shuffled train-test splits, as done in previous studies with these datasets, a trivial feature that encodes temporal autocorrelation not only outperforms LLMs but also accounts for the majority of neural variance that LLMs explain. We therefore use contiguous splits moving forward. Second, we explain the surprisingly high brain scores of untrained LLMs by showing they do not account for additional neural variance beyond two simple features: sentence length and sentence position. This undermines evidence used to claim that the transformer architecture biases computations to be more brain-like. Third, we find that brain scores of trained LLMs on this dataset can largely be explained by sentence length, position, and pronoun-dereferenced static word embeddings; a small, additional amount is explained by sense-specific embeddings and contextual representations of sentence structure. We conclude that over-reliance on brain scores can lead to over-interpretations of similarity between LLMs and brains, and emphasize the importance of deconstructing what LLMs are mapping to in neural signals.

For agents

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

get_harvested_code_for_paper("2406.01538")
get_code_for_paper("2406.01538")
have("2406.01538")

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

curl https://syntology.ai/api/ran/2406.01538.json

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

[![Syntology run record](https://syntology.ai/api/ran/2406.01538.svg)](https://syntology.ai/paper/2406.01538)

Connect an agent — have() is free.