SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2205.12302 · 2022

Garden-Path Traversal in GPT-2

arXiv · PDF · Open in the Atlas

Code that ran

3 of 3 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 0 have no contract verdict; 3 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/2205.12302.json · badge: /api/ran/2205.12302.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
wjurayj/garden-path-gpt2 canonical (classifier label)
classifier basis: classifier_2025
3 ran · 0 no recorded run
FunctionStatusWhere it lives
get_components_nps Ran
ran; our contract draft was wrong, not the code
wjurayj/garden-path-gpt2/make_sents.py
code served (permissive licence) · get_code("414c9ae05b6ae789")
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 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 pandas as pd
import tempfile
import os

# Create a temporary TSV file with sample data
temp_dir = tempfile.mkdtemp()
temp_file = os.path.join(temp_dir, 'nps.tsv')

# Create sample data matching the expected columns
data = {
    'Subject': ['The dog', 'A cat'],
    'Context': ['In the park', 'At home'],
    'Ambiguous Verb': ['saw', 'chased'],
    'Unambiguous Verb': ['observed', 'pursued'],
    'That': ['that', 'that'],
    'NP/S': ['the man', 'the mouse'],
    'Extension': ['running', 'sleeping'],
    'Disambiguator': ['quickly', 'slowly'],
    'Rest': ['away', 'nearby']
}

df = pd.DataFrame(data)
df.to_csv(temp_file, sep='\t', index=False)

# Extract the filename without extension
filename_without_ext = temp_file[:-4]  # Remove '.tsv'

ARGS = {
    'filename': filename_without_ext,
    'sent_type': 'gp',
    'context': False,
    'extension': False,
    'that': False
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
Imports
pandas — 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
get_components_npz Ran
ran; our contract draft was wrong, not the code
wjurayj/garden-path-gpt2/make_sents.py
code served (permissive licence) · get_code("2cb7dc1989f89a4e")
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 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 pandas as pd
import tempfile
import os

# Create a temporary TSV file with the expected structure
temp_dir = tempfile.mkdtemp()
tsv_filename = os.path.join(temp_dir, 'test.tsv')

# Create sample data matching the expected columns
data = {
    'Start': ['The', 'A'],
    'Context': ['in the park', 'by the lake'],
    'Intransitive Verb': ['ran', 'jumped'],
    'Transitive Verb': ['chased', 'caught'],
    'Blocker': ['but', 'and'],
    'Comma': [',', ','],
    'NP/Z': ['dog', 'cat'],
    'Extension': ['quickly', 'suddenly'],
    'Verb': ['barked', 'meowed'],
    'Rest': ['loudly', 'softly']
}

df = pd.DataFrame(data)
df.to_csv(tsv_filename, sep='\t', index=False)

# Extract the filename without extension for the function
filename_without_ext = tsv_filename[:-4]  # Remove '.tsv'

ARGS = {
    'filename': filename_without_ext,
    'sent_type': 'gp',
    'context': False,
    'extension': False,
    'comma': False
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
Imports
pandas — 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
make_sents_npz Ran
ran; our contract draft was wrong, not the code
wjurayj/garden-path-gpt2/make_sents.py
code served (permissive licence) · get_code("5c65d5eab2dcae81")
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 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.

import pandas as pd
import tempfile
import os

# Create a temporary TSV file with the expected structure
temp_dir = tempfile.mkdtemp()
filename = os.path.join(temp_dir, 'npz')

# Create sample data matching the expected columns
data = {
    'Start': ['The', 'A'],
    'Context': ['in the park', 'by the river'],
    'Intransitive Verb': ['ran', 'walked'],
    'Transitive Verb': ['chased', 'saw'],
    'Blocker': ['but', 'and'],
    'Comma': [',', ','],
    'NP/Z': ['dog', 'cat'],
    'Extension': ['quickly', 'slowly'],
    'Verb': ['jumped', 'flew'],
    'Rest': ['over the fence', 'across the sky']
}

df = pd.DataFrame(data)
df.to_csv(filename + '.tsv', sep='\t', index=False)

ARGS = {
    'filename': filename,
    'sent_type': 'gp',
    'context': False,
    'extension': False,
    'comma': False
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
Imports
pandas — 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 3 samples lifted from this paper's linked repositories, 3 carry a function identity and 0 do not. Of the 3 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 3 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

In recent years, large-scale transformer decoders such as the GPT-x family of models have become increasingly popular. Studies examining the behavior of these models tend to focus only on the output of the language modeling head and avoid analysis of the internal states of the transformer decoder. In this study, we present a collection of methods to analyze the hidden states of GPT-2 and use the model's navigation of garden path sentences as a case study. To enable this, we compile the largest currently available dataset of garden path sentences. We show that Manhattan distances and cosine similarities provide more reliable insights compared to established surprisal methods that analyze next-token probabilities computed by a language modeling head. Using these methods, we find that negating tokens have minimal impacts on the model's representations for unambiguous forms of sentences with ambiguity solely over what the object of a verb is, but have a more substantial impact of representations for unambiguous sentences whose ambiguity would stem from the voice of a verb. Further, we find that analyzing the decoder model's hidden states reveals periods of ambiguity that might conclude in a garden path effect but happen not to, whereas surprisal analyses routinely miss this detail.

For agents

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

get_harvested_code_for_paper("2205.12302")
get_code_for_paper("2205.12302")
have("2205.12302")

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

curl https://syntology.ai/api/ran/2205.12302.json

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

[![Syntology run record](https://syntology.ai/api/ran/2205.12302.svg)](https://syntology.ai/paper/2205.12302)

Connect an agent — have() is free.