SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2109.04711 · EMNLP · 2021

Pre-train or Annotate? Domain Adaptation with a Constrained Budget

Wei Xu, Fan Bai, Alan Ritter

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, 2 violated it and 2 have no contract verdict; 1 failure 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/2109.04711.json · badge: /api/ran/2109.04711.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
bflashcp3f/ProcBERT canonical (classifier label)
classifier basis: census
3 ran · 0 no recorded run
allenai/kb canonical (classifier label)
classifier basis: census
2 ran · 0 no recorded run
FunctionStatusWhere it lives
is_chunk_end Ran
ran, violated the contract we drafted
bflashcp3f/ProcBERT/code/utils/conlleval.py
code served (permissive licence) · get_code("e4736e9992d1de49")
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, shape () 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

# Define test arguments for is_chunk_end function
# The function takes two string parameters: prev_tag and tag
# These represent IOBES chunk tags (e.g., "B-PER", "I-LOC", "O")

# Test case 1: (B-PER, I-PER) -> False (continuation of same chunk)
ARGS = {
    "prev_tag": "B-PER",
    "tag": "I-PER"
}
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 bool, 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
is_chunk_start Ran
ran, violated the contract we drafted
bflashcp3f/ProcBERT/code/utils/conlleval.py
code served (permissive licence) · get_code("405fd06f23bf2a87")
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, shape () 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

# Define ARGS with test cases for is_chunk_start function
# The function takes two parameters: prev_tag and tag (both strings)

ARGS = {
    "prev_tag": "B-PER",
    "tag": "I-PER"
}
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 bool, 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
score Ran
ran on a synthesized input
allenai/kb/bin/tacred_scorer.py
code served (permissive licence) · get_code("9cb44938e773d88b")
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 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 sys
from collections import Counter

NO_RELATION = 'no_relation'

# Create simple test data for the score function
# The function expects two lists of equal length containing relation labels

# Create a key (ground truth) list with various relations and no_relation labels
key = [
    'person:knows',
    'person:knows',
    NO_RELATION,
    'org:founded',
    'org:founded',
    NO_RELATION,
    'person:knows',
    'org:based_in'
]

# Create a prediction list with some correct and incorrect predictions
prediction = [
    'person:knows',      # correct
    'org:founded',       # incorrect
    NO_RELATION,         # correct (both no_relation)
    'org:founded',       # correct
    NO_RELATION,         # incorrect (should be org:founded)
    'person:knows',      # incorrect (should be no_relation)
    'person:knows',      # correct
    'org:based_in'       # correct
]

ARGS = {
    'key': key,
    'prediction': prediction,
    'verbose': True
}
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 tuple, ndim 1, dtype family float — 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
split_tag Ran
ran; our contract draft was wrong, not the code
bflashcp3f/ProcBERT/code/utils/conlleval.py
code served (permissive licence) · get_code("f9a22a5ce23105fd")
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)

# Test case 1: Standard IOBES tag with chunk type
chunk_tag_1 = "B-PER"

# Test case 2: O tag (no chunk type)
chunk_tag_2 = "O"

# Test case 3: Other IOBES prefixes
chunk_tag_3 = "I-LOC"

# We'll use the first test case as the argument
ARGS = {
    "chunk_tag": "B-PER"
}
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
unpack_wsd_training_instance Ran
ran on a synthesized input
allenai/kb/kb/wordnet.py
code served (permissive licence) · get_code("9ac58585dd89748b")
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 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.

import torch
import numpy as np

# Construct a valid context for unpack_wsd_training_instance
# context is a list of dicts representing tokens from WSD datasets

context = [
    {
        'token': 'The',
        'lemma': 'the',
        'pos': 'DET',
        'id': 'd001.s001.t001',
        'senses': ['1:09:00::']
    },
    {
        'token': 'hot-dog',
        'lemma': 'hot_dog',
        'pos': 'NOUN',
        'id': 'd001.s001.t002',
        'senses': ['1:13:01::']
    },
    {
        'token': 'is',
        'lemma': 'be',
        'pos': 'VERB',
        'id': 'd001.s001.t003',
        'senses': ['2:42:00::']
    },
    {
        'token': 'delicious',
    }
]

ARGS = {
    'context': context
}
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 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

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, 4 are also present in other papers whose code we hold (1 with at least one paper dated earlier, 0 the same month, 3 later) and 1 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
is_chunk_end 1 other paper
1 later
2310.16427 (2023-10)
PromptAgent: Strategic Planning with Language Models Enables Expert-level Prompt Optimization
is_chunk_start 1 other paper
1 later
2310.16427 (2023-10)
PromptAgent: Strategic Planning with Language Models Enables Expert-level Prompt Optimization
score 1 other paper
1 dated earlier
1906.07510 (2019-06)
Attention Guided Graph Convolutional Networks for Relation Extraction
split_tag 1 other paper
1 later
2310.16427 (2023-10)
PromptAgent: Strategic Planning with Language Models Enables Expert-level Prompt Optimization

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

Recent work has demonstrated that pretraining in-domain language models can boost performance when adapting to a new domain. However, the costs associated with pretraining raise an important question: given a fixed budget, what steps should an NLP practitioner take to maximize performance? In this paper, we view domain adaptation with a constrained budget as a consumer choice problem, where the goal is to select an optimal combination of data annotation and pre-training. We measure annotation costs of three procedural text datasets, along with the pre-training costs of several in-domain language models. The utility of different combinations of pretraining and data annotation are evaluated under varying budget constraints to assess which combination strategy works best. We find that for small budgets, spending all funds on annotation leads to the best performance; once the budget becomes large enough, however, a combination of data annotation and in-domain pre-training yields better performance. Our experiments suggest task-specific data annotation should be part of an economical strategy when adapting an NLP model to a new domain. 1

For agents

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

get_harvested_code_for_paper("2109.04711")
get_code_for_paper("2109.04711")
have("2109.04711")

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

curl https://syntology.ai/api/ran/2109.04711.json

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

[![Syntology run record](https://syntology.ai/api/ran/2109.04711.svg)](https://syntology.ai/paper/2109.04711)

Connect an agent — have() is free.