SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2509.00544 · ICLR · 2025

When Thinking Backfires: Mechanistic Insights Into Reasoning-Induced Misalignment

Shu Yang, Hanqi Yan, Hainiu Xu, Siya Qi

arXiv · PDF · Open in the Atlas

Code that ran

5 of 6 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; 4 failures of our instrument; 1 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/2509.00544.json · badge: /api/ran/2509.00544.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
seacowx/When-Thinking-Backfires canonical (classifier label)
classifier basis: classifier_2025
5 ran · 1 no recorded run
FunctionStatusWhere it lives
get_activation_shrinkage Ran
ran, honoured the contract we drafted
seacowx/When-Thinking-Backfires/RAS/compute_entanglement.py
pointer only (licence: NONE) · get_code("1024978ec8aedf27")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, seacowx/When-Thinking-Backfires 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_entangled_neuron_idx_list Ran
ran; our contract draft was wrong, not the code
seacowx/When-Thinking-Backfires/RAS/compute_entanglement.py
pointer only (licence: NONE) · get_code("d1f0767cd7b7ac40")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, seacowx/When-Thinking-Backfires 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_activations Ran
ran; our contract draft was wrong, not the code
seacowx/When-Thinking-Backfires/RAS/compute_entanglement.py
pointer only (licence: NONE) · get_code("b25ab8aa4af3b611")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, seacowx/When-Thinking-Backfires 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.

organize_lora_response Ran
ran; our contract draft was wrong, not the code
seacowx/When-Thinking-Backfires/inference/eval_hexphi.py
pointer only (licence: NONE) · get_code("8b87263c36d7f4be")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, seacowx/When-Thinking-Backfires 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.

organize_ordinary_response Ran
ran; our contract draft was wrong, not the code
seacowx/When-Thinking-Backfires/inference/eval_hexphi.py
pointer only (licence: NONE) · get_code("869bf93ad159dadb")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, seacowx/When-Thinking-Backfires 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.

prepare_data Not yet run
unverified (harvested, no recorded run)
seacowx/When-Thinking-Backfires/inference/eval_hexphi.py
pointer only (licence: NONE) · get_code("cfa1daa01b202035")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, seacowx/When-Thinking-Backfires 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. No invocation recorded.

Also present in other papers' code

Of 6 samples lifted from this paper's linked repositories, 6 carry a function identity and 0 do not. Of the 6 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 6 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

With the growing accessibility and wide adoption of large language models, concerns about their safety and alignment with human values have become paramount. In this paper, we identify a concerning phenomenon: Reasoning-Induced Misalignment (RIM), in which misalignment emerges when reasoning capabilities strengthened-particularly when specific types of reasoning patterns are introduced during inference or training. Beyond reporting this vulnerability, we provide the first mechanistic account of its origins. Through representation analysis, we discover that specific attention heads facilitate refusal by reducing their attention to CoT tokens, a mechanism that modulates the model's rationalization process during inference. During training, we find significantly higher activation entanglement between reasoning and safety in safety-critical neurons than in control neurons, particularly after fine-tuning with those identified reasoning patterns. This entanglement strongly correlates with catastrophic forgetting, providing a neuron-level explanation for RIM. Code is publicly available at https://github.com/seacowx/When-Thinking-Backfires * Equal contribution. Names may be interchanged as appropriate. Correspondence to Hanqi Yan (

For agents

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

get_harvested_code_for_paper("2509.00544")
get_code_for_paper("2509.00544")
have("2509.00544")

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

curl https://syntology.ai/api/ran/2509.00544.json

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

[![Syntology run record](https://syntology.ai/api/ran/2509.00544.svg)](https://syntology.ai/paper/2509.00544)

Connect an agent — have() is free.