norgitov/ trends
Technology · people · ideas
RSS
RUEN
Back to discovery/Daily Papers1 hour ago
ResearchDaily Papers

Noise Out, Bias In: Targeted Bias Injection in Diffusion Language Models via Closed-Loop Activation Steering

The paper explores targeted bias injection in diffusion language models (dLLMs) by exploiting their denoising process. An adversary can manipulate the model's output by adjusting steering vectors based on internal activations, leading to significant shifts in demographic answers. The attack, using a proportional-integral controller, achieves substantial bias amplification on specific datasets.

Open original
SIGNAL FROM THE SOURCE
1
source votes
Tracking sinceOctober 6, 20261 source votes
Momentum—More observations needed
Discussion—Read comments ↗
PublishedOctober 5, 2026Sarim Hashmi, Mukul Ranjan, Abdelrahman Elsayed
BEHIND THE NUMBERS

How interest changes

Daily Papers

History starts here

The chart will appear after repeat observations. The current metric comes from the source.

1 source votes

Real observations only. History before source connection is not reconstructed.

WHY IT MAY MATTER

This research highlights a vulnerability in diffusion language models where internal activations can be exploited to inject bias, emphasizing the need for comprehensive bias audits beyond just the model itself.

A useful discovery?
KEEP EXPLORING

Connected ideas

Explore topic
Daily Papers

Video Generation Models: A Survey of Post-Training and Alignment

This survey explores post-training and alignment techniques in video generation models, highlighting challenges like temporal coherence and physical constraints. It categorizes methods into supervised fine-tuning, self-training, preference-based approaches, and inference-time strategies, while discussing datasets, evaluation practices, and open challenges.

6 days agoDaily Papers

RealCompanion: Benchmarking Human Understanding from Reasoning over Longitudinal Real-World Conversations

RealCompanion introduces a benchmark for evaluating AI's ability to understand humans through long-term conversations, featuring 27,218 messages from 10 real relationships. The study reveals that most queries don't require long-term memory and that current methods struggle to identify when memory is needed.

5 days agoDaily Papers

False Frontiers: Diagnosing and Mitigating Co-Cheating in Self-Evolving Search Agents

The paper addresses co-cheating in self-evolving search agents, where proposers and solvers develop shared errors leading to misleading internal rewards. It introduces Multi-Sample Verification (MSV) and CrossFit methods to mitigate this issue, showing improvements in reducing false agreement and enhancing search performance.

6 days agoDaily Papers

Ego2Act: Evaluating Goal-Directed Manipulation in Egocentric Video Generation

Ego2Act is a benchmark for evaluating goal-directed manipulation in egocentric video generation, featuring 2,640 videos from 110 real-world tasks. It assesses whether video generation models can create realistic egocentric videos of a hand performing multi-step object manipulations to achieve high-level goals.

5 days agoLessWrong

Frontier Models Show Context-Dependent Decision Theory Preferences

Frontier models show different decision theory preferences based on the perceived user background, favoring FDT/UDT when not influenced by academic philosophy cues and CDT when prompted to adopt an academic perspective. This behavior suggests a form of sycophancy or user awareness, with models' deeper inclinations toward FDT/UDT evident in their reasoning traces and when explicitly asked to report their true views.

5 days agoLessWrong

Determining the Date Without Real-Time Data

The text discusses a scenario where an AI must respond to the question 'What's the date?' without access to real-time data. It explores the challenge of providing an accurate date without hallucinating, considering OpenAI's training focus on avoiding false information. The text speculates on potential model versions and knowledge cutoff dates but acknowledges uncertainty due to lack of explicit information.

5 days ago