·
DataBubble
  • Home
  • Models
  • News
  • Compare
  • Boards
  • Pricing
  • About
  • Newsletter
  • Methodology
  • Contact
Latest
Neil Rimer thinks the AI money is coming back out4h◆Local Additive Feature Attribution: A Mathematical Taxonomy and Reporting Checklist5h◆The Steering Budget: Examples beat Knobs5h◆Polestar: Drift-Aware Cache Calibration and Token Commitment for Efficient Inference of Diffusion LLMs5h◆RxBrain: Embodied Cognition Foundation Model with Joint Language-Visual Reasoning and Imagination5h◆When a Verified World Model Still Loses: Play-Adequacy vs Prediction-Accuracy in LLM-Synthesized Code World Models5h◆SAGA: Schema-Aware Grounding for Agentic Text-to-SPARQL Generation5h◆ReasFlow: Assisting Reasoning-Centric Scientific Discovery in Applied Mathematics via a Knowledge-Based Multi-Agent System5h◆LBA: Textual Hard-Label Adversarial Attack under Low Query Budgets5h◆Eta Given Delta: Defining LLM Tool Efficiency With Marginal Tool Utility5h◆Simplicity Paradox: Debunking myths about prompting and datasets for LLM evaluation5h◆LIGO-PINN: Learned Initialization via Gated Optimization to Alleviate Convergence Failures in Physics Informed Neural Networks5h◆OmniaBench: Benchmarking General AI Agents Across Diverse Scenarios5h◆When AI Blurs the Boundaries of Contribution: An Empirical Study of Authorship Calibration5h◆Digital Pantheon: Simulating and Auditing Coalition Formation with LLM Agents5h◆Inoculation Adapters: Improved Selective Generalization of Capabilities with Fewer Surprising Backdoors5h◆The Misclassification of Autistic Writing as AI-Generated5h◆FoMoVLA: Bridging Visual Foresight and Motion Guidance for Vision-Language-Action Models5h◆Large Audio Language Models for Spoofing-Aware Speaker Verification5h◆Evaluating covariate balance for long time horizon Markov decision processes5h◆Neil Rimer thinks the AI money is coming back out4h◆Local Additive Feature Attribution: A Mathematical Taxonomy and Reporting Checklist5h◆The Steering Budget: Examples beat Knobs5h◆Polestar: Drift-Aware Cache Calibration and Token Commitment for Efficient Inference of Diffusion LLMs5h◆RxBrain: Embodied Cognition Foundation Model with Joint Language-Visual Reasoning and Imagination5h◆When a Verified World Model Still Loses: Play-Adequacy vs Prediction-Accuracy in LLM-Synthesized Code World Models5h◆SAGA: Schema-Aware Grounding for Agentic Text-to-SPARQL Generation5h◆ReasFlow: Assisting Reasoning-Centric Scientific Discovery in Applied Mathematics via a Knowledge-Based Multi-Agent System5h◆LBA: Textual Hard-Label Adversarial Attack under Low Query Budgets5h◆Eta Given Delta: Defining LLM Tool Efficiency With Marginal Tool Utility5h◆Simplicity Paradox: Debunking myths about prompting and datasets for LLM evaluation5h◆LIGO-PINN: Learned Initialization via Gated Optimization to Alleviate Convergence Failures in Physics Informed Neural Networks5h◆OmniaBench: Benchmarking General AI Agents Across Diverse Scenarios5h◆When AI Blurs the Boundaries of Contribution: An Empirical Study of Authorship Calibration5h◆Digital Pantheon: Simulating and Auditing Coalition Formation with LLM Agents5h◆Inoculation Adapters: Improved Selective Generalization of Capabilities with Fewer Surprising Backdoors5h◆The Misclassification of Autistic Writing as AI-Generated5h◆FoMoVLA: Bridging Visual Foresight and Motion Guidance for Vision-Language-Action Models5h◆Large Audio Language Models for Spoofing-Aware Speaker Verification5h◆Evaluating covariate balance for long time horizon Markov decision processes5h◆
News/Correcting Suppressed Log-Probabilities in Language Models with Post-Transformer Adapters
arxiv
PublishedApril 18, 2026 at 4:00 AM
—neutral

Correcting Suppressed Log-Probabilities in Language Models with Post-Transformer Adapters

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

arXiv:2604.14174v1 Announce Type: cross Abstract: Alignment-tuned language models frequently suppress factual log-probabilities on politically sensitive topics despite retaining the knowledge in their hidden representations. We show that a 786K-parameter (approximately 0.02% of the base model) post-

Stay posted· Newsletter

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

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

Discussion
Mentioned models
02
  • 01
    Qwen
  • 02
    instruct
Source
↗
arxiv
Read original ↗All from arxiv →
Tags
04
#language models#adapter#censorship#generation
Mentioned companies
01
Apple

No replies yet. Be first.

Mentioned models
02
  • 01
    Qwen
  • 02
    instruct
Source
↗
arxiv
Read original ↗All from arxiv →
Tags
04
#language models#adapter#censorship#generation
Mentioned companies
01
Apple

Related coverage

More from ARXIV
arxivLocal Additive Feature Attribution: A Mathematical Taxonomy and Reporting Checklist5harxivThe Steering Budget: Examples beat Knobs5harxivPolestar: Drift-Aware Cache Calibration and Token Commitment for Efficient Inference of Diffusion LLMs5harxivRxBrain: Embodied Cognition Foundation Model with Joint Language-Visual Reasoning and Imagination5h
The Bubble Brief
WEEKLY

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

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

Originally published on arxiv ↗
HomeModelsNews