SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2402.09394 · NAACL · 2024

Long-form evaluation of model editing

Frank Rudzicz, Hassan Sajjad, Domenic Rosati, Jinkun Xuemin, Yu Melis, Erkan Yahya, Kayani Satya, Deepika Chavatapalli

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; 1 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/2402.09394.json · badge: /api/ran/2402.09394.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
domenicrosati/longform-evaluation-model-editing canonical (classifier label)
classifier basis: classifier_2025
3 ran · 0 no recorded run
FunctionStatusWhere it lives
calculate_perplexity Ran
ran, honoured the contract we drafted
domenicrosati/longform-evaluation-model-editing/src/automatic_metrics.py
code served (permissive licence) · get_code("6be6c7d3f948ef58")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Sonnet 4.5, prompt python_fixture_v1) and returned float, ndim 0, dtype family float, shape () as a later run recorded it. The image it ran in was not recorded. The code mentions cuda, and the harness aliases cuda to the CPU for every sample; whether this run needed that is 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
import torch.nn as nn

torch.manual_seed(0)

class MockTokenizer:
    def __call__(self, text, return_tensors=None, truncation=False):
        # Return a simple mock tokenized output
        # Simulate tokenizing the passage into a few token IDs
        input_ids = torch.tensor([[101, 2054, 2003, 102]])  # batch_size=1, seq_len=4
        attention_mask = torch.ones_like(input_ids)
        
        class TokenizerOutput:
            def __init__(self, input_ids, attention_mask):
                self.data = {
                    'input_ids': input_ids,
                    'attention_mask': attention_mask
                }
            
            def to(self, device):
                return {k: v.to(device) for k, v in self.data.items()}
            
            def __getitem__(self, key):
                return self.data[key]
        
        return TokenizerOutput(input_ids, attention_mask)

class MockModel(nn.Module):
    def __init__(self):
        super().__init__()
        self.linear = nn.Linear(4, 4)
    
    def forward(self, input_ids, attention_mask=None, labels=None):
        # Return a mock output with a loss attribute
        batch_size, seq_len = input_ids.shape
        
        class ModelOutput:
            def __init__(self, loss):
                self.loss = loss
        
        # Create a simple loss value
        loss = torch.tensor(2.5)
        
        return ModelOutput(loss)

passage = "This is a test passage for perplexity calculation."
model = MockModel()
tokenizer = MockTokenizer()

ARGS = {
    'passage': passage,
    'model': model,
    'tokenizer': tokenizer
}
Setup block drafted by
a language model, Claude Sonnet 4.5, prompt python_fixture_v1
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type float, 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
get_overlap_measures Ran
ran; our contract draft was wrong, not the code
domenicrosati/longform-evaluation-model-editing/src/automatic_metrics.py
code served (permissive licence) · get_code("224fc50d0f23cc6f")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Sonnet 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.

sample = {
    'requested_rewrite': {
        'subject': 'Eiffel Tower',
        'prompt': 'The {} is located in',
        'target_new': {
            'str': ' Berlin'
        },
        'target_true': {
            'str': ' Paris'
        }
    },
    'coupled_prompts_and_properties': {
        'coupled_entities': [
            {
                'entity': 'Arc de Triomphe'
            }
        ]
    },
    'subject_prompt': 'The Eiffel Tower is a famous landmark in France.\nIt was built in 1889.',
    'coupled_prompt': 'The Arc de Triomphe is another famous monument.\nIt stands at the center of Place Charles de Gaulle.'
}

ARGS = {'sample': sample}
Setup block drafted by
a language model, Claude Sonnet 4.5, prompt python_fixture_v1
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 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
sentence_splitter Ran
ran; our contract draft was wrong, not the code
domenicrosati/longform-evaluation-model-editing/src/llm_annotator.py
code served (permissive licence) · get_code("be1f4b21a6bfc222")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Sonnet 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.

text = "This is the first sentence with more than three words. Short one. Another sentence that is long enough to be included. Too short. This final sentence also has enough words to pass the filter."

ARGS = {
    "text": text
}
Setup block drafted by
a language model, Claude Sonnet 4.5, prompt python_fixture_v1
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 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

Evaluations of model editing, a technique for changing the factual knowledge held by Large Language Models (LLMs), currently only use the 'next few token' completions after a prompt. As a result, the impact of these methods on longer natural language generation is largely unknown. We introduce long-form evaluation of model editing (LEME) a novel evaluation protocol that measures the efficacy and impact of model editing in long-form generative settings. Our protocol consists of a machine-rated survey and a classifier which correlates well with human ratings. Importantly, we find that our protocol has very little relationship with previous short-form metrics (despite being designed to extend efficacy, generalization, locality, and portability into a long-form setting), indicating that our method introduces a novel set of dimensions for understanding model editing methods. Using this protocol, we benchmark a number of model editing techniques and present several findings including that, while some methods (ROME and MEMIT) perform well in making consistent edits within a limited scope, they suffer much more from factual drift than other methods. Finally, we present a qualitative analysis that illustrates common failure modes in long-form generative settings including internal consistency, lexical cohesion, and locality issues.

For agents

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

get_harvested_code_for_paper("2402.09394")
get_code_for_paper("2402.09394")
have("2402.09394")

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

curl https://syntology.ai/api/ran/2402.09394.json

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

[![Syntology run record](https://syntology.ai/api/ran/2402.09394.svg)](https://syntology.ai/paper/2402.09394)

Connect an agent — have() is free.