SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2601.08342 · 2026

Detecting Mental Manipulation in Speech via Synthetic Multi-Speaker Dialogue

Wen Liang, Julia Hirschberg, Lin Ai, Ziwei Gong, Run Chen

arXiv · PDF · Open in the Atlas

Code that ran

8 of 11 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 1 has no contract verdict; 6 failures of our instrument; 3 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/2601.08342.json · badge: /api/ran/2601.08342.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
RVC-Boss/GPT-SoVITS canonical (classifier label)
classifier basis: census
8 ran · 3 no recorded run
FunctionStatusWhere it lives
custom_sort_key Ran
ran; our contract draft was wrong, not the code
RVC-Boss/GPT-SoVITS/GPT_SoVITS/inference_webui_fast.py
code served (permissive licence) · get_code("2ba0067e43e2ba5c")
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 re

def custom_sort_key(s):
    parts = re.split('(\\d+)', s)
    parts = [int(part) if part.isdigit() else part for part in parts]
    return parts

# The function takes a single string parameter 's' and applies regex splitting
# followed by type conversion to create a sort key that handles mixed alphanumeric strings

ARGS = {
    's': 'file123name456test'
}
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 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
denorm_spec Ran
ran on a synthesized input
RVC-Boss/GPT-SoVITS/GPT_SoVITS/export_torch_script_v3v4.py
code served (permissive licence) · get_code("cb828fb7e4b52fed")
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 Tensor, ndim 2, dtype family float, shape (2, 4) 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 numpy as np

torch.manual_seed(0)

def denorm_spec(x):
    spec_min = -12
    spec_max = 2
    return (x + 1) / 2 * (spec_max - spec_min) + spec_min

# x should be a tensor that will be normalized
# The function performs denormalization on it
x = torch.randn(2, 4)

ARGS = {
    "x": x
}
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 Tensor, ndim 2, dtype family float, shape (2, 4) — shape (2, 4) 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
inverse_mel_scale Ran
ran; our contract draft was wrong, not the code
RVC-Boss/GPT-SoVITS/GPT_SoVITS/eres2net/kaldi.py
code served (permissive licence) · get_code("cd99ee7d35201131")
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 Tensor, ndim 2, dtype family float, shape (2, 4) 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

torch.manual_seed(0)

# The function takes a mel_freq tensor and applies an inverse mel scale transformation
# It expects a tensor input and returns a tensor output
mel_freq = torch.randn(2, 4)

ARGS = {
    "mel_freq": mel_freq
}
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 Tensor, ndim 2, dtype family float, shape (2, 4) — shape (2, 4) 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
load_sovits_new Ran
ran; our contract draft was wrong, not the code
RVC-Boss/GPT-SoVITS/GPT_SoVITS/export_torch_script.py
code served (permissive licence) · get_code("441dd69fb69b157a")
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 tempfile
import zipfile
from io import BytesIO

torch.manual_seed(0)

# Create a temporary file with a valid PyTorch model saved as a zip (PK header)
with tempfile.NamedTemporaryFile(delete=False, suffix='.pt') as tmp:
    # Create a simple model state dict
    model_state = {
        'layer1.weight': torch.randn(4, 3),
        'layer1.bias': torch.randn(4),
    }
    
    # Save it as a torch file (which uses zip format with PK header)
    torch.save(model_state, tmp.name)
    sovits_path = tmp.name

ARGS = {
    'sovits_path': sovits_path,
}
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
mel_scale Ran
ran; our contract draft was wrong, not the code
RVC-Boss/GPT-SoVITS/GPT_SoVITS/eres2net/kaldi.py
code served (permissive licence) · get_code("d8c64705b3adb44e")
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 Tensor, ndim 2, dtype family float, shape (2, 4) 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
from torch import Tensor

torch.manual_seed(0)

# The mel_scale function takes a frequency tensor and applies a mel-scale transformation
# It performs: 1127.0 * (1.0 + freq / 700.0).log()
# So we need a tensor of frequencies

freq = torch.randn(2, 4)  # Small tensor: batch size 2, 4 frequency values

ARGS = {
    "freq": freq
}
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 Tensor, ndim 2, dtype family float, shape (2, 4) — shape (2, 4) 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
mel_scale_scalar Ran
ran, honoured the contract we drafted
RVC-Boss/GPT-SoVITS/GPT_SoVITS/eres2net/kaldi.py
code served (permissive licence) · get_code("fdeb257eb9797e20")
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

def mel_scale_scalar(freq: float) -> float:
    return 1127.0 * math.log(1.0 + freq / 700.0)

ARGS = {
    "freq": 440.0
}
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 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
norm_spec Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
RVC-Boss/GPT-SoVITS/GPT_SoVITS/export_torch_script_v3v4.py
code served (permissive licence) · get_code("e9f6beb4c18ceca3")
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 Tensor, 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

torch.manual_seed(0)

# The function norm_spec takes a single parameter x (a tensor)
# It normalizes the tensor by shifting and scaling based on spec_min and spec_max
# x should be a tensor with spectrogram values

x = torch.randn(2, 4)  # Small batch of 2 with 4 features

ARGS = {
    "x": x
}
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 Tensor, ndim 2, dtype family float — no output shape is stated: observe_signature.py could not observe it (import:OSError)
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
register_method Ran
ran; our contract draft was wrong, not the code
RVC-Boss/GPT-SoVITS/GPT_SoVITS/TTS_infer_pack/text_segmentation_method.py
code served (permissive licence) · get_code("1ead6905d0f121d5")
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

# The function `register_method` takes a single parameter: `name` (a string)
# It returns a decorator function that takes `func` as a parameter
# The decorator registers the function in a METHODS dict and returns it

# Let's create a simple function to register
def example_function():
    """A simple example function to be registered"""
    return "example"

# Define ARGS mapping the parameter to a value
ARGS = {
    "name": "my_custom_method"
}
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
get_method Not yet run
unverified (harvested, no recorded run)
RVC-Boss/GPT-SoVITS/GPT_SoVITS/TTS_infer_pack/text_segmentation_method.py
code served (permissive licence) · get_code("3144e3365e4a4de1")
How it ran

no invocation recorded.

logits_to_probs Not yet run
unverified (harvested, no recorded run)
RVC-Boss/GPT-SoVITS/GPT_SoVITS/export_torch_script.py
code served (permissive licence) · get_code("cade0a6c392ba656")
How it ran

no invocation recorded.

split_big_text Not yet run
unverified (harvested, no recorded run)
RVC-Boss/GPT-SoVITS/GPT_SoVITS/TTS_infer_pack/text_segmentation_method.py
code served (permissive licence) · get_code("e59729e38275d7ed")
How it ran

no invocation recorded.

Also present in other papers' code

Of 11 samples lifted from this paper's linked repositories, 11 carry a function identity and 0 do not. Of the 11 distinct functions with an identity, 1 are also present in other papers whose code we hold (1 with at least one paper dated earlier, 0 the same month, 0 later) and 10 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
inverse_mel_scale 1 other paper
1 dated earlier
2310.17864 (2023-10)
TorchAudio 2.1: Advancing speech recognition, self-supervised learning, and audio processing components for PyTorch

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

Mental manipulation, the strategic use of language to covertly influence or exploit others, is a newly emerging task in computational social reasoning. Prior work has focused exclusively on textual conversations, overlooking how manipulative tactics manifest in speech. We present the first study of mental manipulation detection in spoken dialogues, introducing a synthetic multi-speaker benchmark SPEECH-MENTALMANIP that augments a text-based dataset with high-quality, voice-consistent Textto-Speech rendered audio. Using few-shot large audio-language models and human annotation, we evaluate how modality affects detection accuracy and perception. Our results reveal that models exhibit high specificity but markedly lower recall on speech compared to text, suggesting sensitivity to missing acoustic or prosodic cues in training. Human raters show similar uncertainty in the audio setting, underscoring the inherent ambiguity of manipulative speech. Together, these findings highlight the need for modality-aware evaluation and safety alignment in multimodal dialogue systems.

For agents

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

get_harvested_code_for_paper("2601.08342")
get_code_for_paper("2601.08342")
have("2601.08342")

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

curl https://syntology.ai/api/ran/2601.08342.json

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

[![Syntology run record](https://syntology.ai/api/ran/2601.08342.svg)](https://syntology.ai/paper/2601.08342)

Connect an agent — have() is free.