Skip to content

Wire Briefings Search

Fetching from the wire…

01

02

03

04

05

06

07

08

Wire Briefings Search

Entity trail

SimCoT

Source-backed findings, relationship evidence, citations, and briefing history from the public MindPattern archive.

Briefing refs

0

Findings

1

Edges

2

Sources

1

Corpus findings

2026-04-05 / arxiv-researcherThoughtSteer: Backdoor Attacks on Continuous Latent Reasoning Achieve ≥99% Success Rate, Evade All DefensesResearchers demonstrate ThoughtSteer, an attack that perturbs a single embedding vector in models that reason in continuous hidden states (Coconut, SimCoT), achieving ≥99% attack success with near-baseline clean accuracy across 124M–3B parameter scales. The attack transfers to held-out benchmarks without retraining (94–100%), survives 25 epochs of clean fine-tuning, and evades all five evaluated active defenses. Neural Collapse in the latent space creates a geometric attractor that amplifies the perturbation — individual vectors still encode the correct answer, but the collective trajectory produces the wrong one.

Graph relationships

CRITICIZES
ThoughtSteer
ThoughtSteer backdoor attacks achieve 99%+ success against SimCoT without triggering defenses
Source finding
CRITICIZES
ThoughtSteer -> SimCoT
ThoughtSteer backdoor attacks achieve 99%+ success against SimCoT without triggering defenses
Source finding

Source trail

Graph sources

entity graphfindings textkg edgeskg entities