arXiv:2609.04678v1 Announce Type: new Abstract: Existing post-training pipelines for coding and terminal agents suffer severe token and control fidelity…
Category: AI
SQL-Zero: Self-Evolving Text-to-SQL
arXiv:2609.04697v1 Announce Type: new Abstract: Training a competitive Text-to-SQL agent usually depends on human-annotated natural-language/SQL pairs,…
ERPBench: Evaluating LLM Agents for Enterprise Decision-Making Across Competitive Market Ecologies
arXiv:2609.04667v1 Announce Type: new Abstract: Large language model (LLM) agents are increasingly proposed for enterprise workflows, yet existing…
Leveraging Imperfect Restoration for Data Availability Attack
arXiv:2609.04627v1 Announce Type: new Abstract: The abundance of online data is at risk of unauthorized usage in training deep learning models. To counter…
Harness-agnostic detection and immunization of reward hacking in self-evolving language models
arXiv:2609.04665v1 Announce Type: new Abstract: Self-evolving language models improve by proposing candidate updates and keeping whatever raises a visible…
A Cost-Aware Agentic Architecture for NL-to-SQL over Nested Enterprise Schemas, with a New Benchmark
arXiv:2609.04641v1 Announce Type: new Abstract: Natural-language-to-SQL systems have ad- vanced rapidly on academic benchmarks, yet production enterprise…
SiLR: Structure-Preserving Admission and Process Reward for LLM Tool Agents
arXiv:2609.04629v1 Announce Type: new Abstract: A runtime gate for an LLM tool agent is usually cast as a filter. In a ReAct loop a rejected proposal is…
NEURA and SECO Partner on Robot Compute Modules Built in Europe
NEURA Robotics and SECO announced a strategic partnership on September 7, 2026, under which the Italian embedded-computing company will design, engineer…
Continual Graph Memory for Adaptive Recommendation under Intent Drift
arXiv:2609.04651v1 Announce Type: new Abstract: This paper studies adaptive recommendation under intent drift, where feedback from each recommendation…
Extremely Sparse Supervision Incentivizes Reasoning Ability
arXiv:2609.04565v1 Announce Type: new Abstract: Large language models demonstrate increasingly strong reasoning capabilities through effective…