SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2003.10687 · EMNLP Findings · 2020

FELIX: Flexible Text Editing Through Tagging and Insertion

Jonathan Mallinson, Aliaksei Severyn, Eric Malmi, Guillermo Garrido

arXiv · PDF · Open in the Atlas

Code that ran

3 of 5 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (3 of them constructed an object rather than computing a result; 0 honoured the contract we drafted, 0 violated it and 3 have no contract verdict; 0 failures of our instrument; 2 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/2003.10687.json · badge: /api/ran/2003.10687.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
phanxuanphucnd/Felix community
Papers with Code archive
3 ran · 1 no recorded run
reichang182/variable-length-piano-infilling community
Papers with Code archive
0 ran · 1 no recorded run
FunctionStatusWhere it lives
Biaffine Ran
ran on a synthesized input · constructed an object; no result was computed
phanxuanphucnd/Felix/tagging/tagging_model.py
code served (permissive licence) · get_code("172a3f793c8e0266")
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 an instance of its own class, Biaffine: the run constructed the object and did not call it, so no result was computed. 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 torch.nn as nn

torch.manual_seed(0)

# Create a Biaffine instance to test
ARGS = {
    'n_in': 4,
    'n_out': 1,
    'scale': 0,
    'bias_x': True,
    'bias_y': True,
}
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 Biaffine — the run constructed an instance of its own class and did not call it; no shape is stated — the instrument coerces any output to an array; for a Biaffine 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
PositionEmbedding Ran
ran on a synthesized input · constructed an object; no result was computed
phanxuanphucnd/Felix/tagging/tagging_model.py
code served (permissive licence) · get_code("18123aa7c54d3cae")
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 an instance of its own class, PositionEmbedding: the run constructed the object and did not call it, so no result was computed. 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 torch.nn as nn
import math

torch.manual_seed(0)

# The PositionEmbedding.__init__ takes embedding_dim, dropout, and max_len
# The forward method expects input x of shape [sequence_length, batch_size, embedding_dim]
# and returns output of the same shape after adding positional embeddings and dropout

# Create a small input tensor
# Shape: [sequence_length=3, batch_size=2, embedding_dim=4]
x = torch.randn(3, 2, 4)

ARGS = {
    'embedding_dim': 4,
    'dropout': 0.1,
    'max_len': 128
}
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 PositionEmbedding — the run constructed an instance of its own class and did not call it; no shape is stated — the instrument coerces any output to an array; for a PositionEmbedding 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
SelfAttentionMask Ran
ran on a synthesized input · constructed an object; no result was computed
phanxuanphucnd/Felix/tagging/tagging_model.py
code served (permissive licence) · get_code("95bd3502053fda29")
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 an instance of its own class, SelfAttentionMask: the run constructed the object and did not call it, so no result was computed. 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 torch.nn as nn

torch.manual_seed(0)

# Set up device (CPU only)
device = 'cpu'

# Create sample tensors for the forward pass
batch_size = 2
from_seq_length = 4
to_seq_length = 3
embedding_dim = 5

# inputs[0]: from_tensor of shape [batch_size, from_seq_length, embedding_dim]
from_tensor = torch.randn(batch_size, from_seq_length, embedding_dim)

# inputs[1]: to_mask of shape [batch_size, to_seq_length]
# The mask should be int32 as per docstring, but the forward method converts to dtype of inputs
to_mask = torch.randint(0, 2, (batch_size, to_seq_length), dtype=torch.int32)

# ARGS dict mapping every parameter of SelfAttentionMask
ARGS = {
    'device': device
}
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 SelfAttentionMask — the run constructed an instance of its own class and did not call it; no shape is stated — the instrument coerces any output to an array; for a SelfAttentionMask 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
FelixTagger Not yet run
unverified (harvested, no recorded run)
phanxuanphucnd/Felix/tagging/tagging_model.py
code served (permissive licence) · get_code("88fb085e02a34db9")
How it ran

no invocation recorded.

prepare_data_for_training Not yet run
unverified (harvested, no recorded run)
reichang182/variable-length-piano-infilling/baselines/FELIX/prepare_data.py
pointer only (licence: GPL-3.0) · get_code("eec13e8669a57c1f")
How it ran

The body of this sample is withheld: GPL-3.0, as GitHub detects it on this paper's repository, reichang182/variable-length-piano-infilling, is not in the permissive set Syntology serves inline. 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. No invocation 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, 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 5 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

We present FELIX -a flexible text-editing approach for generation, designed to derive maximum benefit from the ideas of decoding with bi-directional contexts and self-supervised pretraining. In contrast to conventional sequenceto-sequence (seq2seq) models, FELIX is efficient in low-resource settings and fast at inference time, while being capable of modeling flexible input-output transformations. We achieve this by decomposing the text-editing task into two sub-tasks: tagging to decide on the subset of input tokens and their order in the output text and insertion to in-fill the missing tokens in the output not present in the input. The tagging model employs a novel Pointer mechanism, while the insertion model is based on a Masked Language Model (MLM). Both of these models are chosen to be non-autoregressive to guarantee faster inference. FELIX performs favourably when compared to recent text-editing methods and strong seq2seq baselines when evaluated on four NLG tasks:

For agents

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

get_harvested_code_for_paper("2003.10687")
get_code_for_paper("2003.10687")
have("2003.10687")

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

curl https://syntology.ai/api/ran/2003.10687.json

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

[![Syntology run record](https://syntology.ai/api/ran/2003.10687.svg)](https://syntology.ai/paper/2003.10687)

Connect an agent — have() is free.