·
DataBubble
  • Home
  • Models
  • News
  • Compare
  • Boards
  • Pricing
  • About
  • Newsletter
  • Methodology
  • Contact
Latest
Runtime Authorization for Resources Acquired by AI Agents4h◆Data Attribution via Sketched Metadifferentiation4h◆Do small language models know what they don't know?4h◆Collaborative Memory for Multi-Agent VLM Systems4h◆CaLR: Causal Latent Revision for Robust Diffusion Reasoning4h◆LoRA Enhanced Contrastive Learning with SAS Vision Transformers4h◆Can Agents Design Better Chips with a Higher Level Abstraction?4h◆Talking Past the Machine: Morality, Politeness, and Alignment in Human-AI Dialogue4h◆AtomEgo: Exploring Ego-Robot Integration for Embodied Foundation Model Pretraining4h◆Samsone: A Family of Open Small Audio Language Models for On-Device Inference4h◆On the Limitations of Large Language Models for Conceptual Database Modeling4h◆Intent-Governed Tool Authorization for AI Agents4h◆MOSCOPT: Mixture-of-Skills Collective Optimization for LLM Agents4h◆How a Cooperative-Override Circuit Suppresses Nash Play in Large Language Models4h◆PAVE: Predictive Alignment and Value-Guided Evolution for World-Action Policies4h◆HMB-GAN: Hybrid Multi-B\'ezier GAN for Vector Shape Synthesis4h◆TierKV: Long-Context On-Device LLMs via Predictive Multi-Tier KV Caching4h◆Transsion's Speaker-Attributed Multilingual ASR System for the MLC-SLM 2026 Challenge4h◆RecreationWorld: Scalable and Verifiable Environments for Hybrid Computer-Use Agents4h◆Consolidating RLVR Capabilities Across Domains: A Deep Dive into Fusion Paradigms4h◆Runtime Authorization for Resources Acquired by AI Agents4h◆Data Attribution via Sketched Metadifferentiation4h◆Do small language models know what they don't know?4h◆Collaborative Memory for Multi-Agent VLM Systems4h◆CaLR: Causal Latent Revision for Robust Diffusion Reasoning4h◆LoRA Enhanced Contrastive Learning with SAS Vision Transformers4h◆Can Agents Design Better Chips with a Higher Level Abstraction?4h◆Talking Past the Machine: Morality, Politeness, and Alignment in Human-AI Dialogue4h◆AtomEgo: Exploring Ego-Robot Integration for Embodied Foundation Model Pretraining4h◆Samsone: A Family of Open Small Audio Language Models for On-Device Inference4h◆On the Limitations of Large Language Models for Conceptual Database Modeling4h◆Intent-Governed Tool Authorization for AI Agents4h◆MOSCOPT: Mixture-of-Skills Collective Optimization for LLM Agents4h◆How a Cooperative-Override Circuit Suppresses Nash Play in Large Language Models4h◆PAVE: Predictive Alignment and Value-Guided Evolution for World-Action Policies4h◆HMB-GAN: Hybrid Multi-B\'ezier GAN for Vector Shape Synthesis4h◆TierKV: Long-Context On-Device LLMs via Predictive Multi-Tier KV Caching4h◆Transsion's Speaker-Attributed Multilingual ASR System for the MLC-SLM 2026 Challenge4h◆RecreationWorld: Scalable and Verifiable Environments for Hybrid Computer-Use Agents4h◆Consolidating RLVR Capabilities Across Domains: A Deep Dive into Fusion Paradigms4h◆
News/Bad Genius: Counterfactual-Guided Harness Evolution Beyond Task-Specific Shortcuts
arxiv
PublishedSeptember 21, 2026 at 4:00 AM
—neutral

Bad Genius: Counterfactual-Guided Harness Evolution Beyond Task-Specific Shortcuts

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

arXiv:2609.18366v2 Announce Type: replace Abstract: Reliable agent evaluation is complicated by automatic harness optimization, which repeatedly uses a released benchmark $B_{\mathrm{rel}}$ to guide a Proposer that edits prompts, memory, retrieval, tools, and control code around a fixed target agent

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
arxivRuntime Authorization for Resources Acquired by AI Agents4harxivData Attribution via Sketched Metadifferentiation4harxivDo small language models know what they don't know?4harxivCollaborative Memory for Multi-Agent VLM Systems4h
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 ↗
HomeModelsNews