SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2203.09313 · 2022

EVA2.0: Investigating Open-Domain Chinese Dialogue Systems with Large-Scale Pre-Training

arXiv · PDF · Open in the Atlas

Code that ran

13 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; 2 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/2203.09313.json · badge: /api/ran/2203.09313.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
thu-coai/EVA canonical (classifier label)
classifier basis not recorded on the link
12 ran · 0 no recorded run
MindSpore-scientific/code-6 community (classifier label)
classifier basis not recorded on the link
1 ran · 0 no recorded run
FunctionStatusWhere it lives
add_fp16_config_args Ran
ran on a synthesized input
thu-coai/EVA/src/arguments.py
code served (permissive licence) · get_code("e8c5531fe7efa529")
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 ArgumentParser. 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 argparse
import torch

torch.manual_seed(0)

# Construct a valid ArgumentParser
parser = argparse.ArgumentParser(description='Test parser')

# Define ARGS mapping - the function only takes one parameter
ARGS = {
    'parser': parser
}
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 ArgumentParser — a ArgumentParser output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a ArgumentParser 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
add_model_config_args Ran
ran on a synthesized input
thu-coai/EVA/src/arguments.py
code served (permissive licence) · get_code("41b2dbbded142fc1")
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 ArgumentParser. 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 argparse

def add_model_config_args(parser: argparse.ArgumentParser):
    """Model arguments"""
    group = parser.add_argument_group('model', 'model configuration')
    group.add_argument('--model-config', type=str, help='The path to the model configuration file.')
    group.add_argument('--model-parallel-size', type=int, default=1, help='The segment number of model parallism.')
    return parser

# Create a fresh ArgumentParser instance
parser = argparse.ArgumentParser()

ARGS = {
    'parser': parser
}
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 ArgumentParser — a ArgumentParser output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a ArgumentParser 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
add_training_args Ran
ran on a synthesized input
thu-coai/EVA/src/arguments.py
code served (permissive licence) · get_code("81695f53d395ab64")
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 ArgumentParser. 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 argparse

ARGS = {
    'parser': argparse.ArgumentParser(),
}
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 ArgumentParser — a ArgumentParser output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a ArgumentParser 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
calc_banned_antonym_words_ids Ran
ran on a synthesized input
thu-coai/EVA/src/generation_utils.py
code served (permissive licence) · get_code("114cf78c86343fa4")
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 2, 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)

class MockTokenizer:
    def __init__(self):
        self.vocab = {
            0: '<unk>',
            1: 'hello',
            2: 'world',
            3: 'good',
            4: 'bad',
            5: '<sep>',
            6: 'nice',
            7: 'ugly'
        }
        self.token_to_id = {v: k for k, v in self.vocab.items()}
    
    def convert_ids_to_tokens(self, ids):
        return [self.vocab.get(int(id_val), '<unk>') for id_val in ids]
    
    def convert_tokens_to_ids(self, tokens):
        if isinstance(tokens, str):
            tokens = [tokens]
        return [self.token_to_id.get(token, 0) for token in tokens]

# Create mock tokenizer
tokenizer = MockTokenizer()

# Create input_tokens: batch of 2, sequence length 4
input_tokens = torch.tensor([
    [1, 3, 5, 0],  # "hello good <sep> <unk>"
    [2, 4, 5, 0]   # "world bad <sep> <unk>"
])

# Create antonym_dict mapping words to their antonyms
antonym_dict = {
    'hello': ['goodbye'],
    'world': ['universe'],
    'good': ['bad'],
    'bad': ['good'],
    'nice': ['ugly'],
    'ugly': ['nice'],
    '<sep>': [],
    '<unk>': []
}

ARGS = {
    'input_tokens': input_tokens,
    'tokenizer': tokenizer,
    'antonym_dict': antonym_dict
}
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 2, 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
calculate_gain Ran
ran on a synthesized input
MindSpore-scientific/code-6/neural-audio-synthesis-wavenet/WaveNetEncoder.py
code served (permissive licence) · get_code("72d3cab82bc65d33")
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 float, 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 math
import torch
import torch.nn as nn

torch.manual_seed(0)

# The function takes two parameters:
# - nonlinearity: a string specifying the type of nonlinearity
# - param: an optional parameter (not used in the function body)

ARGS = {
    'nonlinearity': 'relu',
    'param': None
}
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 float, 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
construct_antonym_dict Ran
ran on a synthesized input
thu-coai/EVA/src/generation_utils.py
code served (permissive licence) · get_code("af65f3906cd81298")
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 defaultdict. 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 os
import tempfile
from collections import defaultdict

# Create a temporary directory with the antonym file structure
temp_dir = tempfile.mkdtemp()
antonym_dir = os.path.join(temp_dir, 'antonym')
os.makedirs(antonym_dir, exist_ok=True)

# Create a sample antonym.txt file
antonym_file = os.path.join(antonym_dir, 'antonym.txt')
with open(antonym_file, 'w', encoding='utf-8') as f:
    f.write("('good', 'bad')\n")
    f.write("('hot', 'cold')\n")
    f.write("('big', 'small')\n")

# Create a simple args class
class Args:
    def __init__(self, rule_path):
        self.rule_path = rule_path

ARGS = {
    'args': Args(rule_path=temp_dir)
}
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 defaultdict — a defaultdict output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a defaultdict 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
convert_to_unicode Ran
ran; our contract draft was wrong, not the code
thu-coai/EVA/src/tokenization_eva.py
pointer only (licence: MIT); the stored copy get_code serves, from huawei-noah/Pretrained-Language-Model, is pointer only (licence: NONE) · get_code("1923fc05163d207d")
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 huawei-noah/Pretrained-Language-Model, 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.

get_attn_masks Ran
ran on a synthesized input
thu-coai/EVA/src/eva_interactive.py
code served (permissive licence) · get_code("824e9c994c649f1d")
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

torch.manual_seed(0)

class Args:
    def __init__(self):
        self.fp16 = False

class Tokenizer:
    def __init__(self):
        self.pad_id = 0

# Create small tensors with distinct sizes
batch_size = 2
enc_seq_length = 4
dec_seq_length = 3

contexts = torch.randint(1, 10, (batch_size, enc_seq_length))
contexts[0, 3] = 0  # Add padding to first sample
contexts[1, 2:] = 0  # Add padding to second sample

targets = torch.randint(1, 10, (batch_size, dec_seq_length))

ARGS = {
    'args': Args(),
    'tokenizer': Tokenizer(),
    'contexts': contexts,
    'targets': targets
}
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 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
init_method_normal Ran
ran on a synthesized input
thu-coai/EVA/src/model/eva_modeling.py
code served (permissive licence) · get_code("448c7945683d52e4")
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 function. 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)

# The function init_method_normal takes a std parameter (standard deviation)
# and returns an initializer function that applies normal distribution initialization
# The std parameter should be a float value

ARGS = {
    "std": 0.02
}
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 function — a function output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a function 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
load_vocab Ran
ran; our contract draft was wrong, not the code
thu-coai/EVA/src/tokenization_eva.py
pointer only (licence: MIT); the stored copy get_code serves, from blcuicall/CCL2022-CLTC, is pointer only (licence: NONE) · get_code("deb4d1793cf3ba73")
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 blcuicall/CCL2022-CLTC, 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.

merge Ran
ran on a synthesized input
thu-coai/EVA/src/change_mp.py
code served (permissive licence) · get_code("88a41f3211226c0e")
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

torch.manual_seed(0)

# Create model_parts: a list of state dicts (dictionaries of tensors)
# The function expects to iterate over these and merge them

# First model part
model_part_1 = {
    'project.weight': torch.randn(6, 4),  # shape[0]=6, divisible by 3
    'project_q.weight': torch.randn(8, 4),
    'project_kv.weight': torch.randn(4, 4),  # shape[0]=4, divisible by 2
    'word_embeds.weight': torch.randn(10, 4),
    'dense_relu_dense.wi_1.weight': torch.randn(5, 4),
    'dense_relu_dense.wi_0.weight': torch.randn(7, 4),
    'lm_head.weight': torch.randn(9, 4),
    'other_2d.weight': torch.randn(3, 4),
    'role_embeds.weight': torch.randn(2, 4),
    'bias_1d': torch.randn(4),
}

# Second model part with same keys but different values
model_part_2 = {
    'project.weight': torch.randn(6, 4),
    'project_q.weight': torch.randn(8, 4),
    'project_kv.weight': torch.randn(4, 4),
    'word_embeds.weight': torch.randn(10, 4),
    'dense_relu_dense.wi_1.weight': torch.randn(5, 4),
    'dense_relu_dense.wi_0.weight': torch.randn(7, 4),
    'lm_head.weight': torch.randn(9, 4),
    'other_2d.weight': torch.randn(3, 4),
    'role_embeds.weight': torch.randn(2, 4),
    'bias_1d': torch.randn(4),
}

model_parts = [model_part_1, model_part_2]

ARGS = {
    'model_parts': model_parts,
}
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 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
my_lcs Ran
ran, honoured the contract we drafted
thu-coai/EVA/src/generation_metrics.py
code served (permissive licence) · get_code("1e4734d93a0d4a9e")
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.

# Analyzing the function: my_lcs takes two parameters:
# - string: list of str (tokens from a string split by whitespace)
# - sub: list of str (shorter string tokens split by whitespace)
# The function compares tokens element-wise and builds a 2D DP table

# Create example tokenized strings
string_tokens = ["the", "quick", "brown", "fox", "jumps"]
sub_tokens = ["quick", "fox", "jumps"]

ARGS = {
    "string": string_tokens,
    "sub": sub_tokens,
}
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 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
split Ran
ran on a synthesized input
thu-coai/EVA/src/change_mp.py
code served (permissive licence) · get_code("61d59af2776f4b6e")
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 tqdm

torch.manual_seed(0)

# The function expects:
# - model: a dict where values are tensors (1D or 2D)
# - mp: an integer representing number of model partitions

# Build a small model dict with various tensor keys that match the function's logic
model = {
    'project.weight': torch.randn(12, 8),  # 2D tensor, will be split into 3 parts per partition
    'project_q.weight': torch.randn(8, 6),  # 2D tensor, split evenly
    'project_kv.weight': torch.randn(16, 7),  # 2D tensor, split into 2 parts per partition
    'word_embeds.weight': torch.randn(20, 5),  # 2D tensor, split along dim 0
    'dense_relu_dense.wi_1.weight': torch.randn(12, 4),  # 2D tensor, split along dim 0
    'dense_relu_dense.wi_0.weight': torch.randn(10, 6),  # 2D tensor, split along dim 0
    'lm_head.weight': torch.randn(14, 8),  # 2D tensor, split along dim 0
    'role_embeds.weight': torch.randn(5, 3),  # 2D tensor but has 'role_embeds' in key, not split
    'other_weight': torch.randn(4, 16),  # 2D tensor, split along dim 1
    'bias': torch.randn(7),  # 1D tensor, not split
}

mp = 2  # number of partitions

ARGS = {
    'model': model,
    'mp': mp,
}
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 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

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, 4 are also present in other papers whose code we hold (4 with at least one paper dated earlier, 0 the same month, 2 later) and 9 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
convert_to_unicode 36 other papers
23 dated earlier, 12 later, 1 same year, order not determinable
1609.08144 (2016-09)
Google's Neural Machine Translation System: Bridging the Gap between Human and Machine Translation
1810.04805 (2018-10)
BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding
1901.04085 (2019-01)
Passage Re-ranking with BERT
and 33 more not named here
load_vocab 4 other papers
2 dated earlier, 2 later
1902.10909 (2019-02)
BERT for Joint Intent Classification and Slot Filling
1903.09588 (2019-03)
Utilizing BERT for Aspect-Based Sentiment Analysis via Constructing Auxiliary Sentence
2310.18738 (2023-10)
TLM: Token-Level Masking for Transformers
and 1 more not named here
init_method_normal 2 other papers
1 dated earlier, 1 same year, order not determinable
2109.04332 (2021-09)
PPT: Pre-trained Prompt Tuning for Few-shot Learning
2022.acl-long.576 (2022 (year only))
PPT: Pre-trained Prompt Tuning for Few-shot Learning
my_lcs 1 other paper
1 dated earlier
1909.06356 (2019-09)
Addressing Semantic Drift in Question Generation for Semi-Supervised Question Answering

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

Large-scale pre-training has shown remarkable performance in building open-domain dialogue systems. However, previous works mainly focus on showing and evaluating the conversational performance of the released dialogue model, ignoring the discussion of some key factors towards a powerful human-like chatbot, especially in Chinese scenarios. In this paper, we conduct extensive experiments to investigate these under-explored factors, including data quality control, model architecture designs, training approaches, and decoding strategies. We propose EVA2.0, a large-scale pre-trained open-domain Chinese dialogue model with 2.8 billion parameters, and will make our models and codes publicly available. Automatic and human evaluations show that EVA2.0 significantly outperforms other open-source counterparts. We also discuss the limitations of this work by presenting some failure cases and pose some future research directions on large-scale Chinese open-domain dialogue systems.

For agents

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

get_harvested_code_for_paper("2203.09313")
get_code_for_paper("2203.09313")
have("2203.09313")

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

curl https://syntology.ai/api/ran/2203.09313.json

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

[![Syntology run record](https://syntology.ai/api/ran/2203.09313.svg)](https://syntology.ai/paper/2203.09313)

Connect an agent — have() is free.