·
DataBubble
  • Home
  • Models
  • News
  • Compare
  • Boards
  • Pricing
  • About
  • Newsletter
  • Methodology
  • Contact
Latest
Reasoning Externalization for Faithful Large Language Model Narratives of Stock Return Predictions4h◆Consistent Plan-Act for Long-Horizon Agentic Tasks4h◆Predictive Self-Supervised Learning Provably Identifies Stochastic Signals under Nuisance4h◆Boosting Adversarial Robustness and Generalization with Dictionary Structure4h◆PivotOPD: Learning to Recover from Pivotal Mistakes in Multi-Turn Agents4h◆TED:Text-Axis Evidence Decomposition for Prompted Anomaly Localization4h◆COMiT: Learning Structured Visual Tokens through Sequential Communication4h◆On the (In)effectiveness of AMR Augmentation for Large Language Models4h◆cua-speedrun: Standardized Benchmarking of the Speed of Computer-Use Agents4h◆Compress to Remember: Learning Compact Memory via On-Policy Distillation for Long Video Generation4h◆Building Socio-Affective Artificial Intelligence for Interactive Multi-Agent Simulations4h◆Auditing Agent Actions through Query-Conditioned Attribution4h◆Mitigating Memorization In Language Models4h◆MoEless: Efficient MoE LLM Serving with Serverless Experts4h◆Fork-Think with Confidence4h◆Where Does Randomness Matter in Neural Cellular Automata?4h◆Learning to Route in Visual Space via Multi-Step Embedding Retrieval4h◆Data-Free Pruning of Self-Attention Layers in LLMs4h◆Screening Is Enough4h◆Data Unlearning via Inverse Distillation4h◆Reasoning Externalization for Faithful Large Language Model Narratives of Stock Return Predictions4h◆Consistent Plan-Act for Long-Horizon Agentic Tasks4h◆Predictive Self-Supervised Learning Provably Identifies Stochastic Signals under Nuisance4h◆Boosting Adversarial Robustness and Generalization with Dictionary Structure4h◆PivotOPD: Learning to Recover from Pivotal Mistakes in Multi-Turn Agents4h◆TED:Text-Axis Evidence Decomposition for Prompted Anomaly Localization4h◆COMiT: Learning Structured Visual Tokens through Sequential Communication4h◆On the (In)effectiveness of AMR Augmentation for Large Language Models4h◆cua-speedrun: Standardized Benchmarking of the Speed of Computer-Use Agents4h◆Compress to Remember: Learning Compact Memory via On-Policy Distillation for Long Video Generation4h◆Building Socio-Affective Artificial Intelligence for Interactive Multi-Agent Simulations4h◆Auditing Agent Actions through Query-Conditioned Attribution4h◆Mitigating Memorization In Language Models4h◆MoEless: Efficient MoE LLM Serving with Serverless Experts4h◆Fork-Think with Confidence4h◆Where Does Randomness Matter in Neural Cellular Automata?4h◆Learning to Route in Visual Space via Multi-Step Embedding Retrieval4h◆Data-Free Pruning of Self-Attention Layers in LLMs4h◆Screening Is Enough4h◆Data Unlearning via Inverse Distillation4h◆
News/The Unequal Influence of Bad Advice: Using Training Data Attribution to Modulate Emergent Misalignment
arxiv
PublishedSeptember 30, 2026 at 4:00 AM
—neutral

The Unequal Influence of Bad Advice: Using Training Data Attribution to Modulate Emergent Misalignment

Source
arxiv.orgfull article ↗
Read on arxiv→
Publisher summary· verbatim

arXiv:2609.37914v1 Announce Type: cross Abstract: Fine-tuning large language models on narrow, misaligned tasks can undo their post-training alignment and induce novel misaligned behaviors -- a phenomenon known as \emph{emergent misalignment} (EM). EM has been linked to persona-like representations,

Stay posted· Newsletter

A 5-min weekly brief — top movers, price watch, story of the week.

// no spam · unsubscribe one-click · free forever

Discussion
Source
↗
arxiv
Read original ↗All from arxiv →

No replies yet. Be first.

Source
↗
arxiv
Read original ↗All from arxiv →

Related coverage

More from ARXIV
arxivReasoning Externalization for Faithful Large Language Model Narratives of Stock Return Predictions4harxivConsistent Plan-Act for Long-Horizon Agentic Tasks4harxivPredictive Self-Supervised Learning Provably Identifies Stochastic Signals under Nuisance4harxivBoosting Adversarial Robustness and Generalization with Dictionary Structure4h
The Bubble Brief
WEEKLY

Read AI insights every Tuesday — top movers, new releases, story of the week.

// no spam · unsubscribe one-click · free forever

Originally published on arxiv ↗
Built by Marouane Gazouzi
HomeModelsNews