SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2601.12973 · 2026

Pardon? Evaluating Conversational Repair in Large Audio-Language Models

Wenxuan Zhang, Xuan Zhao, Chong Feng, Yanghao Zhou, Shuanghong Huang, Jinlei Xu, Youchao Zhou

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 (0 of them constructed an object rather than computing a result; 0 honoured the contract we drafted, 0 violated it and 0 have no contract verdict; 2 failures of our instrument; 0 no recorded run), measured 2026-11-22; 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/2601.12973.json · badge: /api/ran/2601.12973.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
sheunghung/EAR canonical (classifier label)
classifier basis: census
2 ran · 0 no recorded run
FunctionStatusWhere it lives
generate_text_only Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
sheunghung/EAR/Eval/baichuanomni.py
code served (permissive licence) · get_code("1a8f3ccbc8ab2fdc")
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. The code mentions cuda, and the harness aliases cuda to the CPU for every sample; whether this run needed that is 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
import re

torch.manual_seed(0)

class MockTokenizer:
    def __init__(self):
        self.vocab_size = 50257
    
    def decode(self, token_ids):
        """Mock decode that returns a string with the token ids"""
        return "Generated text about the topic. <|endoftext|> More text."

class MockProcessor:
    def __init__(self):
        pass
    
    def __call__(self, content):
        """Process content and return mock processor output"""
        output = MockProcessorOutput()
        return output

class MockProcessorOutput:
    def __init__(self):
        self.input_ids = torch.tensor([[1, 2, 3, 4]])
        self.attention_mask = torch.tensor([[1, 1, 1, 1]])
        self.audios = None
        self.encoder_length = None
        self.bridge_length = None

class MockGenerateOutput:
    def __init__(self):
        # Sequences should be longer than input to allow slicing
        self.sequences = torch.tensor([[1, 2, 3, 4, 5, 6, 7, 8, 9, 10]])

class MockModel(nn.Module):
    def __init__(self):
        super().__init__()
        self.processor = MockProcessor()
    
    def generate(self, input_ids, attention_mask=None, audios=None, 
                 encoder_length=None, bridge_length=None, tokenizer=None,
                 max_new_tokens=1024, stop_strings=None, do_sample=True,
                 temperature=0.8, top_k=20, top_p=0.85, repetition_penalty=1.1,
                 return_dict_in_generate=True):
        """Mock generate method"""
        return MockGenerateOutput()

model = MockModel()
tokenizer = MockTokenizer()
content = "This is a test prompt about machine learning."

ARGS = {
    "model": model,
    "tokenizer": tokenizer,
    "content": content
}
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
init_model Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
sheunghung/EAR/Eval/baichuanomni.py
code served (permissive licence) · get_code("37933fa53e4a33fd")
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. The code mentions cuda, and the harness aliases cuda to the CPU for every sample; whether this run needed that is 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
from unittest.mock import MagicMock, Mock
from transformers import PreTrainedModel, PreTrainedTokenizer

torch.manual_seed(0)

# Create a minimal mock tokenizer
class MockTokenizer:
    def __init__(self):
        self.vocab_size = 50257
        self.pad_token_id = 50256
        self.eos_token_id = 50256
        self.bos_token_id = 50256

# Create a minimal mock model that satisfies the function's usage
class MockModel(torch.nn.Module):
    def __init__(self):
        super().__init__()
        self.linear = torch.nn.Linear(4, 4)
        self.training = True
    
    def bind_processor(self, tokenizer, training=False, relative_path='/'):
        # Method that the function calls
        pass
    
    def cuda(self):
        # Return self to allow chaining, but don't actually move to CUDA
        return self
    
    def to(self, *args, **kwargs):
        return self

# Mock the AutoModelForCausalLM.from_pretrained to return our mock model
def mock_from_pretrained(model_path, trust_remote_code=True, torch_dtype=None):
    return MockModel()

def mock_tokenizer_from_pretrained(model_path, trust_remote_code=True):
    return MockTokenizer()

# Patch the transformers functions
import transformers
original_model = transformers.AutoModelForCausalLM.from_pretrained
original_tokenizer = transformers.AutoTokenizer.from_pretrained

transformers.AutoModelForCausalLM.from_pretrained = staticmethod(mock_from_pretrained)
transformers.AutoTokenizer.from_pretrained = staticmethod(mock_tokenizer_from_pretrained)

# Define the arguments
ARGS = {
    "model_path": "gpt2"
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch, transformers — 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

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

Large Audio-Language Models (LALMs) have demonstrated strong performance in spoken question answering (QA), with existing evaluations primarily focusing on answer accuracy and robustness to acoustic perturbations. However, such evaluations implicitly assume that spoken inputs remain semantically answerable, an assumption that often fails in real-world interaction when essential information is missing. In this work, we introduce a repair-aware evaluation setting that explicitly distinguishes between answerable and unanswerable audio inputs. We define answerability as a property of the input itself and construct paired evaluation conditions using a semantic-acoustic masking protocol. Based on this setting, we propose the Evaluability Awareness and Repair (EAR) score, a non-compensatory metric that jointly evaluates task competence under answerable conditions and repair behavior under unanswerable conditions. Experiments on two spoken QA benchmarks across diverse LALMs reveal a consistent gap between answer accuracy and conversational reliability: while many models perform well when inputs are answerable, most fail to recognize semantic unanswerability and initiate appropriate conversational repair. These findings expose a limitation of prevailing accuracy-centric evaluation practices and motivate reliability assessments that treat unanswerable inputs as cues for repair and continued interaction. The core code and dataset are publicly available at https://github.com/sheunghung/EAR.

For agents

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

get_harvested_code_for_paper("2601.12973")
get_code_for_paper("2601.12973")
have("2601.12973")

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

curl https://syntology.ai/api/ran/2601.12973.json

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

[![Syntology run record](https://syntology.ai/api/ran/2601.12973.svg)](https://syntology.ai/paper/2601.12973)

Connect an agent — have() is free.