SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2107.03444 · ACL · 2021

Keep it Simple: Unsupervised Simplification of Multi-Paragraph Text

U Berkeley, St Louis, Philippe Laban, Tobias Schnabel, Marti Hearst, Paul Microsoft, Joe Monahan, Paula Langsam, Camp Leaders

arXiv · PDF · Open in the Atlas

Code that ran

11 of 13 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 10 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/2107.03444.json · badge: /api/ran/2107.03444.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
tingofurro/keep_it_simple canonical (classifier label)
classifier basis: census
11 ran · 2 no recorded run
FunctionStatusWhere it lives
any_gpu_with_space Ran
ran on a synthesized input
tingofurro/keep_it_simple/utils_misc.py
code served (permissive licence) · get_code("3ee08534dd7b401f")
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 bool, ndim 0, 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.

import numpy as np
import os
import time
import tqdm

# Mock the nvidia-smi output by creating a temporary file with GPU memory info
def setup_mock_gpu_memory():
    """Create a mock tmp_smi file for testing"""
    mock_output = """GPU 0
    FB Memory Usage
        Total                             : 16384 MB
        Reserved                          : 8192 MB
        Allocated                         : 4096 MB
        Free                              : 4096 MB
GPU 1
    FB Memory Usage
        Total                             : 16384 MB
        Reserved                          : 8192 MB
        Allocated                         : 2048 MB
        Free                              : 6144 MB
"""
    with open('tmp_smi', 'w') as f:
        f.write(mock_output)

# Setup mock GPU memory file before defining ARGS
setup_mock_gpu_memory()

ARGS = {
    'gb_needed': 2.0
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, tqdm — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type bool, ndim 0, dtype family float — no output shape is stated: observe_signature.py could not observe it (fixture:OSError)
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
cc_news_collate Ran
ran on a synthesized input
tingofurro/keep_it_simple/train_keep_it_simple.py
code served (permissive licence) · get_code("248292fc5ffe5c2b")
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 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.

import torch

torch.manual_seed(0)

# The function expects inps to be an iterable of dicts
# Each dict should have a 'text' key with string value
# The function splits by '\n', counts spaces, and sorts paragraphs

inps = [
    {
        'text': 'short\nthis is a paragraph with many words in it to test\nanother line here'
    },
    {
        'text': 'one\ntwo three four five six seven eight nine ten eleven twelve thirteen\nshort'
    }
]

ARGS = {
    'inps': inps
}
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 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
fix_punctuation Ran
ran on a synthesized input
tingofurro/keep_it_simple/utils_edits.py
code served (permissive licence) · get_code("211650defcae2522")
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

ARGS = {
    "text": "Hello , world . I can't believe it 's working . I 'm happy and we 're done"
}
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
get_constant_schedule Ran
ran on a synthesized input
tingofurro/keep_it_simple/utils_optim.py
code served (permissive licence) · get_code("4575f85aa7122560")
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 LambdaLR. 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
from torch.optim import SGD

torch.manual_seed(0)

# Create a simple model
model = nn.Linear(4, 2)

# Create an optimizer
optimizer = SGD(model.parameters(), lr=0.1)

# Define arguments for get_constant_schedule
ARGS = {
    "optimizer": optimizer,
    "last_epoch": -1
}
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 LambdaLR — a LambdaLR output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a LambdaLR 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
get_constant_schedule_with_warmup Ran
ran on a synthesized input
tingofurro/keep_it_simple/utils_optim.py
code served (permissive licence) · get_code("a2003a42a1bf5632")
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 LambdaLR. 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
from torch.optim import SGD

torch.manual_seed(0)

# Create a simple model
model = nn.Linear(4, 3)

# Create an optimizer
optimizer = SGD(model.parameters(), lr=0.1)

# Define arguments
ARGS = {
    "optimizer": optimizer,
    "num_warmup_steps": 100,
    "last_epoch": -1
}
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 LambdaLR — a LambdaLR output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a LambdaLR 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
get_pos_group Ran
ran on a synthesized input
tingofurro/keep_it_simple/utils_masking.py
code served (permissive licence) · get_code("a25190d7f66f8704")
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 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.

import torch

torch.manual_seed(0)

ARGS = {
    'idx': 1
}
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 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
mask_tokens Ran
ran on a synthesized input
tingofurro/keep_it_simple/utils_masking.py
code served (permissive licence) · get_code("aae67c85186c7e28")
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 tuple, ndim 2, dtype family int; 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.

import torch
import numpy as np

torch.manual_seed(0)

# Analyze the function:
# - tokens: iterable of values that can be compared with 'in' and 'not in'
# - kw_idxs: container that supports 'in' membership testing
# - mask_idx: scalar value (default 0), used as replacement for masked tokens

# Create simple test data
tokens = [0, 1, 2, 3, 1, 2, 0]  # list of token IDs
kw_idxs = {1, 2}  # set of keyword indices to mask
mask_idx = 0  # replacement value for masked tokens

ARGS = {
    "tokens": tokens,
    "kw_idxs": kw_idxs,
    "mask_idx": mask_idx,
}
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 tuple, ndim 2, dtype family int — a tuple 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
pad Ran
ran, honoured the contract we drafted
tingofurro/keep_it_simple/model_generator.py
code served (permissive licence) · get_code("aa982ce4f9f219e4")
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 Tensor, ndim 3, dtype family float, shape (2, 5, 4) as a later run recorded it. The image it ran in was 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)

# pad() takes two parameters: data and padval
# data should be a sequence of tensors (used by pad_sequence)
# padval is a scalar padding value (default 0)

# Create a list of tensors with different lengths to test padding
data = [
    torch.randn(3, 4),  # 3 timesteps, 4 features
    torch.randn(5, 4),  # 5 timesteps, 4 features
]

padval = 0

ARGS = {
    "data": data,
    "padval": padval,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch, tqdm — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Tensor, ndim 3, dtype family float, shape (2, 5, 4) — shape (2, 5, 4) 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
select_logprobs Ran
ran on a synthesized input
tingofurro/keep_it_simple/utils_rl.py
code served (permissive licence) · get_code("f019c240efcb2696")
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 Tensor, ndim 1, dtype family float, shape (2,) as a later run recorded it. The image it ran in was 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 numpy as np

torch.manual_seed(0)

# Create logits tensor: shape (batch_size, seq_length, vocab_size)
# batch_size=2, seq_length=5, vocab_size=4
logits = torch.randn(2, 5, 4)

# Create decoded_tokens: list of lists, one per batch item
# Each inner list contains token indices (must be valid indices into vocab_size=4)
decoded_tokens = [
    [0, 1, 2],           # First sequence
    [1, 2, 0, 1]         # Second sequence
]

# eos_id: end-of-sequence token id
eos_id = 3

ARGS = {
    "logits": logits,
    "decoded_tokens": decoded_tokens,
    "eos_id": eos_id
}
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 Tensor, ndim 1, dtype family float, shape (2,) — shape (2,) 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
show_diff Ran
ran on a synthesized input
tingofurro/keep_it_simple/utils_edits.py
code served (permissive licence) · get_code("910784711d3445e5")
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 Levenshtein
from colorama import Fore

# The function show_diff takes two string parameters:
# - original: the original string
# - modified: the modified string
# It compares them using Levenshtein.opcodes() and returns a colored diff string

ARGS = {
    "original": "hello world",
    "modified": "hallo warld"
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
Levenshtein, colorama — 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
unfold Ran
ran on a synthesized input
tingofurro/keep_it_simple/model_salience.py
code served (permissive licence) · get_code("21cbeb28c1f4b1f1")
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 Tensor, ndim 1, dtype family int, shape (9,) as a later run recorded it. The image it ran in was 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 numpy as np

torch.manual_seed(0)

# The function takes sent_toks (list of sentences, each sentence is a list of tokens)
# and make_tensor (bool flag)
# It flattens the nested list and optionally converts to LongTensor

# Create a small nested list of token indices (simulating tokenized sentences)
sent_toks = [
    [1, 2, 3],
    [4, 5],
    [6, 7, 8, 9]
]

make_tensor = True

ARGS = {
    "sent_toks": sent_toks,
    "make_tensor": make_tensor
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
nltk, numpy, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Tensor, ndim 1, dtype family int, shape (9,) — shape (9,) 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
build_levenshtein_matrix Not yet run
unverified (harvested, no recorded run)
tingofurro/keep_it_simple/utils_edits.py
code served (permissive licence) · get_code("d584842a9e9b1e1c")
How it ran

no invocation recorded.

doc2ents Not yet run
unverified (harvested, no recorded run)
tingofurro/keep_it_simple/model_guardrails.py
code served (permissive licence) · get_code("a51a3b067e73c1d5")
How it ran

no invocation recorded.

Also present in other papers' code

Of 13 samples lifted from this paper's linked repositories, 13 carry a function identity and 0 do not. Of the 13 distinct functions with an identity, 3 are also present in other papers whose code we hold (2 with at least one paper dated earlier, 0 the same month, 3 later) and 10 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
get_constant_schedule_with_warmup 3 other papers
2 dated earlier, 1 later
2104.14757 (2021-04)
An Adversarial Transfer Network for Knowledge Representation Learning
2106.11310 (2021-06)
Towards Long-Form Video Understanding
2204.07275 (2022-04)
Incremental Prompting: Episodic Memory Prompt for Lifelong Event Detection
get_constant_schedule 2 other papers
1 dated earlier, 1 later
2106.11310 (2021-06)
Towards Long-Form Video Understanding
2204.07275 (2022-04)
Incremental Prompting: Episodic Memory Prompt for Lifelong Event Detection
pad 1 other paper
1 later
2405.10260 (2024-05)
Keep It Private: Unsupervised Privatization of Online Text

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

This work presents Keep it Simple (KiS), a new approach to unsupervised text simplification which learns to balance a reward across three properties: fluency, salience and simplicity. We train the model with a novel algorithm to optimize the reward (k-SCST), in which the model proposes several candidate simplifications, computes each candidate's reward, and encourages candidates that outperform the mean reward. Finally, we propose a realistic text comprehension task as an evaluation method for text simplification. When tested on the English news domain, the KiS model outperforms strong supervised baselines by more than 4 SARI points, and can help people complete a comprehension task an average of 18% faster while retaining accuracy, when compared to the original text.

For agents

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

get_harvested_code_for_paper("2107.03444")
get_code_for_paper("2107.03444")
have("2107.03444")

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

curl https://syntology.ai/api/ran/2107.03444.json

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

[![Syntology run record](https://syntology.ai/api/ran/2107.03444.svg)](https://syntology.ai/paper/2107.03444)

Connect an agent — have() is free.