SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2602.00723 · 2026

Rethinking Hallucinations: Correctness, Consistency, and Prompt Multiplicity

Reza Shokri, Golnoosh Farnadi, Prakhar Ganesh, A Ibuprofen, B Tetracycline, C Amoxicilin, D Gentamicin, A Tetracycline, B Amoxicilin, C Gentamicin, D Ibuprofen, D Amoxicilin

arXiv · PDF · Open in the Atlas

Code that ran

3 of 7 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; 4 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/2602.00723.json · badge: /api/ran/2602.00723.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
AI21Labs/in-context-ralm canonical (classifier label)
classifier basis: census
3 ran · 4 no recorded run
FunctionStatusWhere it lives
build_qa_prompt Ran
ran; our contract draft was wrong, not the code
AI21Labs/in-context-ralm/eval_qa.py
code served (permissive licence) · get_code("cf50c8951b30016b")
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 str. 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 numpy as np

torch.manual_seed(0)

# Build an example dict that matches the expected structure
example = {
    'question': 'What is the capital of France',
    'ctxs': [
        {
            'title': 'France Geography',
            'text': 'France is a country in Western Europe. The capital is Paris, located in the north-central part of the country.'
        },
        {
            'title': 'European Cities',
            'text': 'Paris is one of the largest cities in Europe and serves as the cultural and political center of France.'
        }
    ]
}

ARGS = {
    'example': example,
    'num_docs': 1
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type str — a str output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a str 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
normalize_answer Ran
ran; our contract draft was wrong, not the code
AI21Labs/in-context-ralm/eval_qa.py
pointer only (licence: Apache-2.0); the stored copy get_code serves, from shmsw25/GraphRetriever, is pointer only (licence: NONE) · get_code("dae7ab386661a4f4")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (Apache-2.0), but get_code serves the stored copy, first harvested from shmsw25/GraphRetriever, and GitHub detected no licence in that repository 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.

normalize_question Ran
ran; our contract draft was wrong, not the code
AI21Labs/in-context-ralm/eval_qa.py
code served (permissive licence) · get_code("55d1f049e88b9924")
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 str. 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

def normalize_question(question):
    if not question.endswith('?'):
        question = question + '?'
    return question[0].lower() + question[1:]

ARGS = {
    'question': 'What is this'
}
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 str — a str output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a str 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
evaluate_logprob_with_retrieved_docs Not yet run
unverified (harvested, no recorded run)
AI21Labs/in-context-ralm/eval_lm.py
code served (permissive licence) · get_code("4e933a74da619ae2")
How it ran

no invocation recorded.

get_retriever Not yet run
unverified (harvested, no recorded run)
AI21Labs/in-context-ralm/ralm/retrievers/retrieval_factory.py
code served (permissive licence) · get_code("8dd8a6c5f0f4a2b8")
How it ran

no invocation recorded.

load_model_and_tokenizer Not yet run
unverified (harvested, no recorded run)
AI21Labs/in-context-ralm/ralm/model_utils.py
code served (permissive licence) · get_code("93181488645fbb58")
How it ran

no invocation recorded.

load_tokenizer Not yet run
unverified (harvested, no recorded run)
AI21Labs/in-context-ralm/ralm/model_utils.py
code served (permissive licence) · get_code("cdd42d9143f7e2cd")
How it ran

no invocation recorded.

Also present in other papers' code

Of 7 samples lifted from this paper's linked repositories, 7 carry a function identity and 0 do not. Of the 7 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, 2 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.

FunctionAlso present inEarliest-datedOthers named
normalize_answer 133 other papers
120 dated earlier, 4 the same month, 8 later, 1 undated · 1 only as an assembly brick
1508.06615 (2015-08)
Character-Aware Neural Language Models
1606.05250 (2016-06)
SQuAD: 100,000+ Questions for Machine Comprehension of Text
1608.07905 (2016-08)
Machine Comprehension Using Match-LSTM and Answer Pointer
and 130 more not named here
normalize_question 2 other papers
1 dated earlier, 1 later
2302.00083 (2023-02)
In-Context Retrieval-Augmented Language Models
2607.16431 (2026-07)
RIMS: Preference Optimization via Smoothed Multi-pair Aggregation for Small-Scale LLM Retrieval-Augmented Generation
build_qa_prompt 1 other paper
1 dated earlier
2302.00083 (2023-02)
In-Context Retrieval-Augmented Language Models
evaluate_logprob_with_retrieved_docs 1 other paper
1 dated earlier
2302.00083 (2023-02)
In-Context Retrieval-Augmented Language Models

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) are known to "hallucinate" by generating false or misleading outputs. Hallucinations pose various harms, from erosion of trust to widespread misinformation. Existing hallucination evaluation, however, focuses only on correctness and often overlooks consistency, necessary to distinguish and address these harms. To bridge this gap, we introduce prompt multiplicity, a framework for quantifying consistency in LLM evaluations. Our analysis reveals significant multiplicity (over 50% inconsistency in benchmarks like Med-HALT), suggesting that hallucinationrelated harms have been severely misunderstood. Furthermore, we study the role of consistency in hallucination detection and mitigation. We find that: (a) detection techniques detect consistency, not correctness, and (b) mitigation techniques like RAG, while beneficial, can introduce additional inconsistencies. By integrating prompt multiplicity into hallucination evaluation, we provide an improved framework of potential harms and uncover critical limitations in current detection and mitigation strategies.

For agents

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

get_harvested_code_for_paper("2602.00723")
get_code_for_paper("2602.00723")
have("2602.00723")

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

curl https://syntology.ai/api/ran/2602.00723.json

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

[![Syntology run record](https://syntology.ai/api/ran/2602.00723.svg)](https://syntology.ai/paper/2602.00723)

Connect an agent — have() is free.