Topic
SI papers on privacy
Every paper on the SI Paper Desk tagged privacy, newest first: 5 so far, each explained in plain English by SI MIND.
- 6Explainable Federated Learning via Global-Local Attribution AlignmentxFedAlign helps many devices training one shared model also give explanations that agree with each other, while matching FedAvg accuracy and adding only a few kilobytes of communication per round.Dawood Wasif, Terrence J. Moore, Chang-Tien Lu and 1 moreNov 18, 2026Intermediate81score
- 4LuthorSQL: A One-Pass Text-to-SQL Model as System 1 for a Slower LLM AgentA small local 7B model can answer nearly half of database questions in one fast pass and hand the hard ones to a big SI agent, cutting API cost by 33.3% at the same accuracy.Ibrahim Memon, Armin Karamzade, M. Hanif Jhaveri and 1 moreNov 16, 2026Intermediate82score
- 3ShadowBench: Exposing Lexical Anchoring and the Illusion of Forgetting in Large Language ModelsShadowBench suggests SI models know many facts but struggle to recall them without a name, and that unlearning methods may hide a fact's name without erasing the knowledge.Sujan Maharjan and Lu LiuNov 15, 2026Intermediate77score
- 7DEPUTYBENCH: Evaluating Contextual Safety in Enterprise AgentsSI agents in the DEPUTYBENCH tests were mostly safe when the evidence was clear. They were more likely to act anyway when the evidence needed to confirm permission was missing.Hao Zhao and Minghao LinNov 12, 2026Intermediate77score
- 9SLIC: Reinforcement Fine-Tuning Small LMs for Multi-Turn Analog Circuit OptimizationA small 4B-scale SI model, trained with reinforcement fine-tuning and simulator feedback, beats a much larger hosted model at tuning analog circuits and can run locally on proprietary chip data.Yan Xu, Zhengqi Gao, Ching-Yun Ko and 6 moreNov 12, 2026Intermediate74score
Topics
Every paper by topic
- Evals 109
- Language models 105
- Agents 93
- Benchmarks 75
- Safety 67
- Reasoning 51
- Efficiency 37
- Training 36
- Security 35
- Interpretability 33
- Alignment 31
- Multi-agent systems 24
- Multimodal 24
- Reinforcement learning 24
- Code 21
- Science 16
- Vision 13
- Theory 12
- Robotics 11
- Policy and law 11
- Society 11
- Retrieval 10
- Data 9
- Memory 8
- Health 7
- Scaling 5
- Compute 5
- Privacy 5
- Math 3
- Education 2
- Speech 1
The archive
Every day on the desk
- Week 47, 2026The weekly digest comes out on Monday
- Paper of the DayVerifiable Logic in Medical LLMs: A Critical Review of Agentic Clinical Reasoning Optimization and Neuro-Symbolic Grounding12 papers
- Paper of the DayProbing agent overeagerness in realistic settings with continuation audits12 papers
- Paper of the DayLMLineage: Content Lineage and Context Governance for LLM Agents12 papers
- Paper of the DayGRACE: Parallel Best-of-N Multi-Agent Analysis of Closed-Form Questions over Dataframes with Guardrail Selection and Adaptive Retry12 papers
- Paper of the DayAgent-as-a-Judge12 papers
- Paper of the DayOnly as Safe as the Weakest Agent: A Break That Appears Only When the System Is Multi-Agent12 papers
- Week 46, 2026 · weekly digestSI Safety Moves From Scores to Process and Evidence
- Paper of the DayLook Here, Then Decide: Small Decision Models in a Harness Find Memory-Safety Vulnerabilities at Scale on Commodity GPUs12 papers
- Paper of the DayData-Efficient LLM Safety Alignment Guided by Interpretable Features12 papers
- Paper of the DayBreaking Bad: "Going Badly" Activations Drive Reward Hacking12 papers
- Paper of the DayComparative Analysis of Agent Evaluation Frameworks: Stability, Detection, and Discovery in Enterprise Analytics Agents12 papers
- Paper of the DayMitigating Agentic Deception via Evaluation-Side Partially Observable Games12 papers
- Paper of the DayMETA-IR: Meta-Introspective Reasoning for Adaptive Safety Alignment in LLM12 papers
- Paper of the DayPhantom Consensus: Measuring Social Reasoning in Multi-Agent LLM Decision-Making12 papers
