SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2205.12673 · EMNLP · 2022

InstructDial: Improving Zero and Few-shot Generalization in Dialogue through Instruction Tuning

Jeffrey Bigham, Shikib Mehri, Cathy Jiao, Prakhar Gupta, Yi-Ting Yeh, Maxine Eskenazi

arXiv · PDF · Open in the Atlas

Code that ran

2 of 2 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (2 of them constructed an object rather than computing a result; 0 honoured the contract we drafted, 0 violated it and 2 have no contract verdict; 0 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/2205.12673.json · badge: /api/ran/2205.12673.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
prakharguptaz/Instructdial marked official in Papers with Code
Papers with Code archive
2 ran · 0 no recorded run
FunctionStatusWhere it lives
Generator Ran
ran on a synthesized input · constructed an object; no result was computed
prakharguptaz/Instructdial/instruction_files/response_generation.py
code served (permissive licence) · get_code("30835eda93d5d7b2")
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, Generator: 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 random
from collections import defaultdict
import tempfile
import os

# Create minimal stand-in classes
class MinimalDatasetReader:
    def __init__(self, dataset_name, num_examples=3):
        self.dataset_name = dataset_name
        self.split = 'train'
        self.idx = 0
        self.num_examples = num_examples
        self.examples = [
            {
                'dataset': dataset_name,
                'text': f'example {i}',
                'context': ['utterance 1', 'utterance 2'],
                'response': f'response {i}'
            }
            for i in range(num_examples)
        ]
    
    def get_next(self):
        if self.idx >= len(self.examples):
            return None
        dp = self.examples[self.idx]
        self.idx += 1
        return dp

class MinimalArgs:
    def __init__(self):
        self.max_data = 10
        self.data_sample_type = 'common'
        self.tasks_output_folder = tempfile.gettempdir()

class MinimalTaskConfig:
    def __init__(self):
        self.task_name = 'dialogue_response'
        self.max_data = 5

class MinimalSettings:
    MAX_CONTEXT_NUMUTTERANCE = 20
    MAX_DIALOGUE_LENGTH = 400

# Create instances
args = MinimalArgs()

taskconfig = {
    'task_name': 'dialogue_response',
    'max_data': 5
}

data_readers = [
    MinimalDatasetReader('dataset1', num_examples=3),
    MinimalDatasetReader('dataset2', num_examples=2),
]

# Create settings module-like object
settings = MinimalSettings()

ARGS = {
    'args': args,
    'taskconfig': taskconfig,
    'data_readers': data_readers,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
tqdm — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Generator — 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 Generator output the recorded ndim 1 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
GeneratorBasic Ran
ran on a synthesized input · constructed an object; no result was computed
prakharguptaz/Instructdial/instruction_files/response_generation.py
code served (permissive licence) · get_code("093b8e9465b23ba1")
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, GeneratorBasic: 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 random
import json
from collections import Counter, defaultdict
import tempfile
import os

# Set random seed for reproducibility
random.seed(42)

# Create a temporary directory for output
temp_dir = tempfile.mkdtemp()

# Stand-in class for args
class Args:
    def __init__(self):
        self.max_data = 10
        self.tasks_output_folder = temp_dir
        self.data_sample_type = 'individual'  # Can be 'individual', 'common', or other

# Stand-in class for taskconfig
class TaskConfig(dict):
    def __init__(self):
        super().__init__()
        self['task_name'] = 'test_task'
        self['max_data'] = 5

# Create sample data readers (list of data sequences)
data_readers = [
    {
        'dataset': 'opendialkg',
        'context': ['Do you know Manchester United F.C.?', 'Is that a soccer team?'],
        'response': 'Yes, it is a soccer team.'
    },
    {
        'dataset': 'opendialkg',
        'context': ['What do you like?', 'I like sports.'],
        'response': 'That is cool!'
    },
    {
        'dataset': 'convai2',
        'context': ['Yo!', 'yo wassup!'],
        'response': 'How are you doing?'
    },
    {
        'dataset': 'convai2',
        'context': ['Hello', 'Hi there'],
        'response': 'Nice to meet you!'
    },
    {
        'dataset': 'opendialkg',
        'context': ['What is your name?'],
        'response': 'My name is Assistant.'
    }
]

# Create args and taskconfig instances
args = Args()
taskconfig = TaskConfig()

ARGS = {
    'args': args,
    'taskconfig': taskconfig,
    'data_readers': data_readers
}
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 GeneratorBasic — 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 GeneratorBasic output the recorded ndim 1 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 2 samples lifted from this paper's linked repositories, 2 carry a function identity and 0 do not. Of the 2 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 2 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

Instruction tuning is an emergent paradigm in NLP wherein natural language instructions are leveraged with language models to induce zeroshot performance on unseen tasks. Dialogue is an especially interesting area in which to explore instruction tuning because dialogue systems perform multiple tasks related to language (e.g., natural language understanding and generation, domain-specific interaction), yet instruction tuning has not been systematically explored for dialogue-related tasks. We introduce INSTRUCTDIAL, an instruction tuning framework for dialogue, which consists of a repository of 48 diverse dialogue tasks in a unified text-to-text format created from 59 openly available dialogue datasets. We explore crosstask generalization ability on models tuned on INSTRUCTDIAL across diverse dialogue tasks. Our analysis reveals that INSTRUCTDIAL enables good zero-shot performance on unseen datasets and tasks such as dialogue evaluation and intent detection, and even better performance in a few-shot setting. To ensure that models adhere to instructions, we introduce novel meta-tasks. We establish benchmark zero-shot and few-shot performance of models trained using the proposed framework on multiple dialogue tasks 1 .

For agents

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

get_harvested_code_for_paper("2205.12673")
get_code_for_paper("2205.12673")
have("2205.12673")

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

curl https://syntology.ai/api/ran/2205.12673.json

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

[![Syntology run record](https://syntology.ai/api/ran/2205.12673.svg)](https://syntology.ai/paper/2205.12673)

Connect an agent — have() is free.