SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2305.01711 · NeurIPS · 2023

Don't Stop Pretraining? Make Prompt-based Fine-tuning Powerful Learner

Zhengxiang Shi, Aldo Lipani

arXiv · PDF · Open in the Atlas

Code that ran

8 of 8 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; 7 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/2305.01711.json · badge: /api/ran/2305.01711.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
ZhengxiangShi/PowerfulPromptFT canonical (classifier label)
classifier basis: census
8 ran · 0 no recorded run
FunctionStatusWhere it lives
count_parameters Ran
ran, honoured the contract we drafted
ZhengxiangShi/PowerfulPromptFT/src/model.py
code served (permissive licence) · get_code("5f5e5de530a5a8b4")
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 int, ndim 0, dtype family int, 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
import torch.nn as nn

torch.manual_seed(0)

class SimpleModel(nn.Module):
    def __init__(self):
        super().__init__()
        self.linear1 = nn.Linear(4, 3)
        self.linear2 = nn.Linear(3, 2)
        self.frozen_layer = nn.Linear(2, 1)
        # Freeze one layer to have non-trainable parameters
        for param in self.frozen_layer.parameters():
            param.requires_grad = False
    
    def forward(self, x):
        x = self.linear1(x)
        x = self.linear2(x)
        x = self.frozen_layer(x)
        return x

model = SimpleModel()

ARGS = {
    'model': model
}
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 int, ndim 0, dtype family int, 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
dart_template Ran
ran; our contract draft was wrong, not the code
ZhengxiangShi/PowerfulPromptFT/src/convert_to_pretrain_pcp.py
code served (permissive licence) · get_code("ff1816cf44827a08")
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
import torch.nn as nn

torch.manual_seed(0)

# Analyze the function: dart_template(text_a, text_b, label)
# The function takes three string parameters and returns a list
# No complex objects, models, or tensors are used - just simple string concatenation

ARGS = {
    "text_a": "The quick brown fox",
    "text_b": "jumps over the lazy dog",
    "label": "positive"
}
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
dart_ts_template Ran
ran; our contract draft was wrong, not the code
ZhengxiangShi/PowerfulPromptFT/src/convert_to_pretrain_pcp.py
code served (permissive licence) · get_code("d9f08b1debac0613")
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
import numpy as np

torch.manual_seed(0)

ARGS = {
    "text_a": "This is the first text.",
    "text_b": "This is the second text.",
    "label": "positive"
}
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 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
get_label Ran
ran; our contract draft was wrong, not the code
ZhengxiangShi/PowerfulPromptFT/src/convert_to_pretrain_tapt_1.py
pointer only (licence: MIT); the stored copy get_code serves, from alrope123/prompt-waywardness, is pointer only (licence: NONE) · get_code("126ae8eeb63d59bf")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (MIT), but get_code serves the stored copy, first harvested from alrope123/prompt-waywardness, 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.

load_datasets Ran
ran; our contract draft was wrong, not the code
ZhengxiangShi/PowerfulPromptFT/src/convert_to_pretrain_tapt_1.py
pointer only (licence: MIT); the stored copy get_code serves, from uijeongjang/lora-ntk, is pointer only (licence: NONE) · get_code("8e27652a29c8adcc")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (MIT), but get_code serves the stored copy, first harvested from uijeongjang/lora-ntk, 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.

split_header Ran
ran; our contract draft was wrong, not the code
ZhengxiangShi/PowerfulPromptFT/src/convert_to_pretrain_tapt_1.py
pointer only (licence: MIT); the stored copy get_code serves, from alrope123/prompt-waywardness, is pointer only (licence: NONE) · get_code("efd1b7718b950470")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (MIT), but get_code serves the stored copy, first harvested from alrope123/prompt-waywardness, 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.

ssl_template Ran
ran; our contract draft was wrong, not the code
ZhengxiangShi/PowerfulPromptFT/src/convert_to_pretrain_pcp.py
code served (permissive licence) · get_code("edb11b8546838e1f")
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

def ssl_template(text_a: str, text_b: str, label: str):
    return [text_a, text_b, 'It was', label, '.']

ARGS = {
    'text_a': 'The quick brown fox',
    'text_b': 'jumps over the lazy dog',
    'label': 'positive'
}
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
text_classification_metrics Ran
ran; our contract draft was wrong, not the code
ZhengxiangShi/PowerfulPromptFT/src/processors.py
code served (permissive licence) · get_code("b62076c5483738dc")
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

torch.manual_seed(0)

# Construct arguments for text_classification_metrics
task_name = "sentiment"

# Create predictions and labels as numpy arrays
# The function compares them with == and calls .mean()
preds = np.array([0, 1, 0, 1], dtype=np.int64)
labels = np.array([0, 1, 1, 1], dtype=np.int64)

ARGS = {
    "task_name": task_name,
    "preds": preds,
    "labels": labels,
}
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 8 samples lifted from this paper's linked repositories, 8 carry a function identity and 0 do not. Of the 8 distinct functions with an identity, 8 are also present in other papers whose code we hold (4 with at least one paper dated earlier, 0 the same month, 8 later) and 0 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
split_header 7 other papers
4 dated earlier, 3 later
2108.04106 (2021-08)
Noisy Channel Language Model Prompting for Few-Shot Text Classification
2112.08348 (2021-12)
Prompt Waywardness: The Curious Case of Discretized Interpretation of Continuous Prompts
2211.03044 (2022-11)
Tuning Language Models as Training Data Generators for Augmentation-Enhanced Few-Shot Learning
and 4 more not named here
text_classification_metrics 6 other papers
4 dated earlier, 2 later
2110.01256 (2021-10)
Revisiting Self-Training for Few-Shot Learning of Language Model
2110.06274 (2021-10)
LiST: Lite Prompted Self-training Makes Parameter-Efficient Few-shot Learners
2205.01308 (2022-05)
Contrastive Learning for Prompt-Based Few-Shot Language Learners
and 3 more not named here
get_label 5 other papers
3 dated earlier, 2 later
2108.04106 (2021-08)
Noisy Channel Language Model Prompting for Few-Shot Text Classification
2112.08348 (2021-12)
Prompt Waywardness: The Curious Case of Discretized Interpretation of Continuous Prompts
2211.03044 (2022-11)
Tuning Language Models as Training Data Generators for Augmentation-Enhanced Few-Shot Learning
and 2 more not named here
load_datasets 4 other papers
2 dated earlier, 2 later
2108.04106 (2021-08)
Noisy Channel Language Model Prompting for Few-Shot Text Classification
2211.03044 (2022-11)
Tuning Language Models as Training Data Generators for Augmentation-Enhanced Few-Shot Learning
2309.05173 (2023-09)
DePT: Decomposed Prompt Tuning for Parameter-Efficient Fine-tuning
and 1 more not named here
count_parameters 1 other paper
1 later
2309.05173 (2023-09)
DePT: Decomposed Prompt Tuning for Parameter-Efficient Fine-tuning
dart_template 1 other paper
1 later
2309.05173 (2023-09)
DePT: Decomposed Prompt Tuning for Parameter-Efficient Fine-tuning
dart_ts_template 1 other paper
1 later
2309.05173 (2023-09)
DePT: Decomposed Prompt Tuning for Parameter-Efficient Fine-tuning
ssl_template 1 other paper
1 later
2309.05173 (2023-09)
DePT: Decomposed Prompt Tuning for Parameter-Efficient Fine-tuning

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

Language models (LMs) trained on vast quantities of unlabelled data have greatly advanced the field of natural language processing (NLP). In this study, we re-visit the widely accepted notion in NLP that continued pre-training LMs on task-related texts improves the performance of fine-tuning (FT) in downstream tasks. Through experiments on eight single-sentence tasks and eight sentence-pair tasks in both semi-supervised and fully-supervised settings, we find that conventional continued pre-training does not consistently provide benefits and can even be detrimental for sentence-pair tasks or when prompt-based FT is used. To tackle these issues, we propose Prompt-based Continued Pre-training (PCP), which combines the idea of instruction tuning with conventional continued pre-training. Our approach aims to improve the performance of prompt-based FT by presenting both taskrelated texts and prompt templates to LMs through unsupervised pre-training objectives before fine-tuning for the target task. Our empirical evaluations on 21 benchmarks demonstrate that the PCP consistently improves the performance of state-of-the-art prompt-based FT approaches (up to 20.1% absolute) in both semisupervised and fully-supervised settings, even with only hundreds of unlabelled examples. Additionally, prompt-based FT with the PCP outperforms state-of-theart semi-supervised approaches with greater simplicity, eliminating the need for an iterative process and extra data augmentation. Our further analysis explores the performance lower bound of the PCP and reveals that the advantages of PCP persist across different sizes of models and datasets. Code is available at https://github.com/ZhengxiangShi/PowerfulPromptFT.

For agents

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

get_harvested_code_for_paper("2305.01711")
get_code_for_paper("2305.01711")
have("2305.01711")

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

curl https://syntology.ai/api/ran/2305.01711.json

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

[![Syntology run record](https://syntology.ai/api/ran/2305.01711.svg)](https://syntology.ai/paper/2305.01711)

Connect an agent — have() is free.