SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2403.18802 · NeurIPS · 2024

Long-form factuality in large language models

Quoc Le, Dustin Tran, Daiyi Peng, Yifeng Lu, Chengrun Yang, Ruibo Liu, Da Huang, Jie Huang, Xinying Song, Jerry Wei, Nathan Hu, Cosmo Du

arXiv · PDF · Open in the Atlas

Code that ran

5 of 5 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 1 has 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/2403.18802.json · badge: /api/ran/2403.18802.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
google-deepmind/long-form-factuality canonical (classifier label)
classifier basis: census
4 ran · 0 no recorded run
chandralegend/saf-eval community (classifier label)
classifier basis not recorded on the link
1 ran · 0 no recorded run
FunctionStatusWhere it lives
extract_first_square_brackets Ran
ran; our contract draft was wrong, not the code
google-deepmind/long-form-factuality/common/utils.py
pointer only (licence: NOASSERTION) · get_code("7d8151eb9e8c2000")
How it ran

The body of this sample is withheld: GitHub could not classify the licence of this paper's repository, google-deepmind/long-form-factuality (NOASSERTION). The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

get_atomic_annotations Ran
ran; our contract draft was wrong, not the code
google-deepmind/long-form-factuality/eval/correlation_vs_factscore.py
pointer only (licence: NOASSERTION) · get_code("713c896c9eab7aaa")
How it ran

The body of this sample is withheld: GitHub could not classify the licence of this paper's repository, google-deepmind/long-form-factuality (NOASSERTION). The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

get_logger Ran
ran on a synthesized input
chandralegend/saf-eval/saf_eval/utils/logging.py
code served (permissive licence) · get_code("297aaefc12e89e13")
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 SafEvalLogger. 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 logging
import os
import sys
from pathlib import Path

# Define ARGS dict with all parameters for get_logger function
ARGS = {
    'name': 'test-logger',
    'level': logging.INFO,
    'log_dir': None,
    'console': True,
    'file': False,
    'json_format': False,
}
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 SafEvalLogger — a SafEvalLogger output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a SafEvalLogger 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
random_selection Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
google-deepmind/long-form-factuality/common/utils.py
pointer only (licence: NOASSERTION) · get_code("bee924b7c1a1b868")
How it ran

The body of this sample is withheld: GitHub could not classify the licence of this paper's repository, google-deepmind/long-form-factuality (NOASSERTION). The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

strip_string Ran
ran; our contract draft was wrong, not the code
google-deepmind/long-form-factuality/common/utils.py
pointer only (licence: NONE); the stored copy get_code serves, from mbzuai-nlp/fire, is pointer only (licence: NONE) · get_code("0ee0530df323c190")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, google-deepmind/long-form-factuality when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

Also present in other papers' code

Of 5 samples lifted from this paper's linked repositories, 5 carry a function identity and 0 do not. Of the 5 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 4 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
strip_string 1 other paper
1 later
2411.00784 (2024-11)
FIRE: Fact-checking with Iterative Retrieval and Verification

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

Large language models (LLMs) often generate content that contains factual errors when responding to fact-seeking prompts on open-ended topics. To benchmark a model's long-form factuality in open domains, we first use GPT-4 to generate LongFact, a prompt set comprising thousands of questions spanning 38 topics. We then propose that LLM agents can be used as automated evaluators for longform factuality through a method which we call Search-Augmented Factuality Evaluator (SAFE). SAFE utilizes an LLM to break down a long-form response into a set of individual facts and to evaluate the accuracy of each fact using a multi-step reasoning process comprising sending search queries to Google Search and determining whether a fact is supported by the search results. Furthermore, we propose extending F1 score as an aggregated metric for long-form factuality. To do so, we balance the percentage of supported facts in a response (precision) with the percentage of provided facts relative to a hyperparameter representing a user's preferred response length (recall). Empirically, we demonstrate that LLM agents can outperform crowdsourced human annotators-on a set of ∼16k individual facts, SAFE agrees with crowdsourced human annotators 72% of the time, and on a random subset of 100 disagreement cases, SAFE wins 76% of the time. At the same time, SAFE is more than 20 times cheaper than human annotators. We also benchmark thirteen language models on LongFact across four model families (Gemini, GPT, Claude, and PaLM-2), finding that larger language models generally achieve better long-form factuality. LongFact, SAFE, and all experimental code are available at https://github. com/google-deepmind/long-form-factuality.

For agents

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

get_harvested_code_for_paper("2403.18802")
get_code_for_paper("2403.18802")
have("2403.18802")

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

curl https://syntology.ai/api/ran/2403.18802.json

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

[![Syntology run record](https://syntology.ai/api/ran/2403.18802.svg)](https://syntology.ai/paper/2403.18802)

Connect an agent — have() is free.