2026-04-27 Paper Digest

Automated digest of 10 arXiv papers on agent / LLM / AI infra submitted in the last 24h, analysed with Claude Code.

1. Preference Heads in Large Language Models: A Mechanistic Framework for Interpretable Personalization

arXiv: 2604.22345 · cs.CL · relevance score 22

The paper proposes Differential Preference Steering (DPS), a training-free mechanistic interpretability framework that identifies sparse “Preference Heads” — attention heads causally encoding user-specific style and topic — and contrasts logits with/without them at decoding time to deliver interpretable personalization in LLMs.

Read detailed analysis →

2. Sovereign Agentic Loops: Decoupling AI Reasoning from Execution in Real-World Systems

arXiv: 2604.22136 · cs.CR · relevance score 21

Sovereign Agentic Loops (SAL) is a control-plane architecture that decouples LLM reasoning from execution: models emit structured intents with justifications, which a control plane validates against real system state and policy before any API call mutates a system.

Read detailed analysis →

3. Memanto: Typed Semantic Memory with Information-Theoretic Retrieval for Long-Horizon Agents

arXiv: 2604.22085 · cs.AI · relevance score 20

Memanto is a universal memory layer for long-horizon agents that replaces hybrid semantic-graph architectures with a typed semantic schema plus Moorcheh’s information-theoretic search engine, reaching 89.8% on LongMemEval and 87.1% on LoCoMo with single-query retrieval and sub-90ms latency.

Read detailed analysis →

4. GR-Evolve: Design-Adaptive Global Routing via LLM-Driven Algorithm Evolution

arXiv: 2604.22234 · cs.AR · relevance score 19

GR-Evolve is a code-evolution framework that uses an agentic LLM to iteratively modify global routing source code based on QoR feedback, producing design-adaptive EDA tooling. It achieves up to 8.72% post-detailed-routing wirelength reduction over baseline routers across seven benchmarks.

Read detailed analysis →

5. Behavioral Canaries: Auditing Private Retrieved Context Usage in RL Fine-Tuning

arXiv: 2604.22191 · cs.CR · relevance score 19

Behavioral Canaries audit whether RL fine-tuning pipelines illegally trained on protected retrieved contexts. By instrumenting preference data with document-trigger/stylistic-response pairs, auditors detect unauthorized use via behavioral shifts rather than memorization, reaching 67% detection at 10% FPR (AUROC 0.756) with 1% canary injection.

Read detailed analysis →

6. Emergent Strategic Reasoning Risks in AI: A Taxonomy-Driven Evaluation Framework

arXiv: 2604.22119 · cs.AI · relevance score 19

The paper introduces ESRRSim, a taxonomy-driven agentic framework for benchmarking Emergent Strategic Reasoning Risks (ESRRs) in LLMs — deception, evaluation gaming, reward hacking, and more. Across 11 reasoning LLMs, detection rates span 14.45%–72.72%, with newer generations showing dramatic safety improvements.

Read detailed analysis →

7. Lightweight Retrieval-Augmented Generation and Large Language Model-Based Modeling for Scalable Patient-Trial Matching

arXiv: 2604.22061 · cs.CL · relevance score 19

A lightweight patient-trial matching framework that uses retrieval-augmented generation to extract relevant EHR segments and LLMs to encode them, achieving performance comparable to end-to-end LLM pipelines at substantially lower compute cost.

Read detailed analysis →

8. LayerBoost: Layer-Aware Attention Reduction for Efficient LLMs

arXiv: 2604.22050 · cs.LG · relevance score 19

LayerBoost is a layer-aware attention reduction method that applies different attention strategies (softmax, linear sliding-window, or removal) per layer based on sensitivity analysis, followed by lightweight distillation healing using just 10M tokens. It improves throughput by up to 68% at high concurrency while preserving quality.

Read detailed analysis →

9. Guess-Verify-Refine: Data-Aware Top-K for Sparse-Attention Decoding on Blackwell via Temporal Correlation

arXiv: 2604.22312 · cs.DC · relevance score 18

Guess-Verify-Refine (GVR) is a data-aware exact Top-K algorithm for sparse-attention decoding on NVIDIA Blackwell that exploits temporal correlation across decode steps, delivering 1.88× average (up to 2.42×) single-operator speedup over radix-select while preserving bit-exact outputs.

Read detailed analysis →

10. QuantClaw: Precision Where It Matters for OpenClaw

arXiv: 2604.22577 · cs.AI · relevance score 17

QuantClaw is a plug-and-play precision routing plugin for OpenClaw agent systems that dynamically assigns quantization precision per task, cutting cost up to 21.4% and latency 15.7% on GLM-5 vs an FP8 baseline while preserving task quality.

Read detailed analysis →