arXiv subjects

Mike Banad

Publications and source records attributed to Mike Banad.

3 recordsLinked to original sources

arXiv2026

Continuous Kinetic Model of Counterion-Mediated Redox Switching in 14-Bit Molecular Memristors for Analog In-Memory Computing

Kinetic molecular memristors offer a compelling path toward high-density analog in-memory computing by achieving 14-bit precision, wherein discrete redox transitions and counterion displacements dictate non-volatile conductance across a deterministic energy landscape. However, integrating these nanoscale kinetics into macroscale circuit simulators has remained bottlenecked by derivative discontinuities and solver convergence failures. Here, we present a physics-inspired continuous kinetic model that bridges microscopic molecular transport with scalable circuit architectures. By casting discrete intramolecular states, counterion relaxation dynamics, and Marcus-Hush-Chidsey electron transfer rates into continuously differentiable state variables, we derive a robust, Spectre-compatible compact model and a matching vectorized Python solver. This work culminates in a unified computational platform to reliably predict the behavior of complex circuits built on these molecular memristors.

cs.AR↗

arXiv2026

Improving Heart-Focused Medical Question Answering in LLMs via Variance-Aware Rubric Rewards with GRPO

Large language models can support medical question answering, but privacy, cost, and reliability still limit their clinical use in focused domains. Can rubric-based reinforcement learning bring a locally deployable model close to much larger systems? We study Group Relative Policy Optimization (GRPO) for heart-focused medical question answering with rubric supervision from RaR-Medicine. Rather than rely on binary rubric aggregation or a single overall score, our variance-aware reward framework builds continuous rewards from criterion-level outcomes. It preserves partial credit and reduces reward sparsity. We adapt a Qwen3-14B policy with low-rank adapters and quantized weights. On a held-out heart-related HealthBench subset, the best GRPO variant improves accuracy from 0.362 to 0.502 and F1 from 0.532 to 0.668 over the base model, close to GPT-OSS-120B. An independent clinician rated most responses clinically acceptable (61%) and patient ready (63%), both above the automated accuracy; fewer than 2% were unsafe. Taken together, the results support rubric-based reinforcement learning as a practical route for focused medical language models.

cs.CL↗

arXiv2026

Validation-Gated Causal Interventions for Interpreting High-Stakes Large Language Model Behavior: A Case Study in Suicidality Detection

Large language models are proposed to flag suicidal content, but representing a concept and acting on it are distinct. A suicidality feature is decodable from 0.5B parameters upward, while the behavior that uses it emerges only in the low billions. Where it emerges, it rests on a compact mid-network feature: in Llama-3.1-8B-Instruct this direction decodes suicidal from non-suicidal text above a matched lexical baseline and survives removal of the 38 most label-predictive keywords. Projecting it out drops the model's ranking from AUC 0.975 to 0.832, while an equal-norm random direction does not; ablating a low-dimensional subspace degrades it further, though no stable rank is recoverable. The direction transfers to two further corpora, replicates in Qwen and Mistral, and tracks clinician-assigned risk. Steering raises an off-target readout as much as the target across five concepts, so sufficiency claims from yes/no readouts are confounded by generic salience.

cs.CL↗