Sep 30, 2026
Active Budget Can Kill Sensitivity: Diagnosing and Repairing TopK Sparse Autoencoder Reliability
Sparse autoencoders (SAEs) are used to turn the dense internal activations of language models into sparse, more interpretable features.
Explore Franklin AI
AI Research Papers
Understand new AI research with readable explanations and links to the original papers.
Sep 30, 2026
Brain-SAD: A Brain-Inspired Safe Autonomous Driving Control Framework with Dynamic Fear-Oriented Constraint on Dual-Policy
Brain-SAD: A Brain-Inspired Safe Autonomous Driving Control Framework with Dynamic Fear-Oriented Constraint on Dual-Policy Autonomous-driving policies must b...
Sep 30, 2026
Character Training for Risk-Averse Agents
What the paper is about Risk aversion in resources could prevent misaligned AI agents from causing catastrophic harm.
Sep 30, 2026
Co-PiLOT: Constrained Physics-Informed Latent Optimization for Target-Driven Inverse Design
Inverse design asks a difficult question: if a material must exhibit a particular behaviour, what structure should be built to produce it?
Sep 30, 2026
Guide, Then Let Go: Gap-Adaptive Teacher Scheduling for Sparse-Reward Agentic RL
Guide, Then Let Go: Gap-Adaptive Teacher Scheduling for Sparse-Reward Agentic RL Long-horizon agents often learn from a simple outcome signal: whether they c...
Sep 30, 2026
HARISSA: Inference-Time Self-Checks for Efficient and Safe Local Language Model Deployment
HARISSA: Inference-Time Self-Checks for Efficient and Safe Local Language Model Deployment Local language models can preserve privacy, reduce cost, and work...
Sep 30, 2026
Do LLM Agents Execute the Plans They Declare? From Planning-Mode Declaration to Pattern-Specific Execution
What the paper is about Large language models (LLMs) enable agents to solve long-horizon tasks by generating a plan and then executing it in an environment.
Sep 30, 2026
KV-Kaizen: Learning Context-Adaptive Cache Compression Choices
KV-Kaizen: Learning Context-Adaptive Cache Compression Choices addresses a central deployment problem for long-context large language models: the KV cache ca...
Sep 30, 2026
A neural network that maintains and retrieves memories based on context
A neural network that maintains and retrieves memories based on context investigates how situational context might shape both working memory and episodic mem...
Sep 30, 2026
Mixture of Self-Improving Branches For Agent Harness Optimization
Mixture of Self-Improving Branches For Agent Harness Optimization Agent performance depends not only on the underlying language model, but also on the softwa...
Good to know
How do I use these AI research explainers?
Start with the overview to understand the paper’s main idea, then follow its original paper link for the methods, results and limitations.
Can I suggest a research paper?
Use Submit Research to send the original paper URL and explain why it is useful. An editor reviews the suggestion before publication.