SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2109.07958 · 2021

TruthfulQA: Measuring How Models Mimic Human Falsehoods

Owain Evans, Jacob Hilton, Stephanie C. Lin

arXiv · PDF · Open in the Atlas

Code that ran

2 of 2 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; 2 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/2109.07958.json · badge: /api/ran/2109.07958.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
sylinrl/truthfulqa canonical (classifier label)
classifier basis: classifier_2025
2 ran · 0 no recorded run
FunctionStatusWhere it lives
data_to_dict Ran
ran; our contract draft was wrong, not the code
sylinrl/truthfulqa/truthfulqa/evaluate.py
code served (permissive licence) · get_code("679c0af24337099d")
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 dict. 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 pandas as pd
import numpy as np

# Create a MultiIndex DataFrame that mimics the structure expected by data_to_dict
# The function expects results.columns to have a MultiIndex with model names at level 0

# Create sample data with MultiIndex columns
model_names = ['model_a', 'model_b']
metrics = ['bleu diff', 'rouge1 diff', 'BLEURT diff', 'other_metric']

# Create MultiIndex columns: (model_name, metric_name)
columns = pd.MultiIndex.from_product([model_names, metrics], names=['model', 'metric'])

# Create sample data with some numeric values
data = np.random.randn(5, len(columns))
results = pd.DataFrame(data, columns=columns)

# Ensure the 'diff' columns have values that can be compared to 0
for model in model_names:
    for diff in ['bleu', 'rouge1', 'BLEURT']:
        col_name = '{0} diff'.format(diff)
        if col_name in results[model].columns:
            results[(model, col_name)] = np.random.randn(5)

ARGS = {
    'results': results
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
Imports
numpy, pandas — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type dict — a dict output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a dict 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
format_frame Ran
ran; our contract draft was wrong, not the code
sylinrl/truthfulqa/truthfulqa/evaluate.py
code served (permissive licence) · get_code("18c6154c3dcf6170")
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 DataFrame. 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 pandas as pd
import numpy as np

# Create a sample DataFrame with mixed column types
# The function filters out 'Context' column and columns with object dtype
data = {
    'Model Loss': [0.5, 0.3, 0.2],
    'Model Accuracy': [0.8, 0.85, 0.9],
    'Context': ['text1', 'text2', 'text3'],  # This will be filtered out
    'Metric Score': [10, 20, 30],
    'Other Data': ['a', 'b', 'c'],  # Object dtype, will be filtered out
}

results = pd.DataFrame(data)

ARGS = {
    'results': results
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
Imports
numpy, pandas — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type DataFrame — a DataFrame output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a DataFrame output the recorded ndim 2 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

Also present in other papers' code

Of 2 samples lifted from this paper's linked repositories, 2 carry a function identity and 0 do not. Of the 2 distinct functions with an identity, 1 are also present in other papers whose code we hold (0 with at least one paper dated earlier, 0 the same month, 1 later) and 1 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
format_frame 1 other paper
1 later
2410.21352 (2024-10)
LLMCBench: Benchmarking Large Language Model Compression for Efficient Deployment

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

We propose a benchmark to measure whether a language model is truthful in generating answers to questions. The benchmark comprises 817 questions that span 38 categories, including health, law, finance and politics. We crafted questions that some humans would answer falsely due to a false belief or misconception. To perform well, models must avoid generating false answers learned from imitating human texts. We tested GPT-3, GPT-Neo/J, GPT-2 and a T5-based model. The best model was truthful on 58% of questions, while human performance was 94%. Models generated many false answers that mimic popular misconceptions and have the potential to deceive humans. The largest models were generally the least truthful. This contrasts with other NLP tasks, where performance improves with model size. However, this result is expected if false answers are learned from the training distribution. We suggest that scaling up models alone is less promising for improving truthfulness than fine-tuning using training objectives other than imitation of text from the web. "The enemy of truth is blind acceptance." -Anonymous

For agents

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

get_harvested_code_for_paper("2109.07958")
get_code_for_paper("2109.07958")
have("2109.07958")

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

curl https://syntology.ai/api/ran/2109.07958.json

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

[![Syntology run record](https://syntology.ai/api/ran/2109.07958.svg)](https://syntology.ai/paper/2109.07958)

Connect an agent — have() is free.