Results for “scaling-laws”

Scaling Data-Constrained Language Models

Training Compute-Optimal Large Language Models

Scaling Instruction-Finetuned Language Models

Configure and troubleshoot Salesforce Case Escalation Rules, including time-based entries, business hours, and escalation actions, with guidance on diagnosing why cases are not escalating.

Apply Fitts's Law to size and position interactive targets for fast, accurate interaction.

Design scalable distributed systems using structured approaches for load balancing, caching, database scaling, and message queues.

Type Scale

Guides scaling decisions for Qdrant vector databases when data volume exceeds single-node capacity, covering tenant scaling, time window rotation, vertical scaling, and horizontal sharding.

Guides scaling Qdrant vector search with time-based data rotation using shard rotation, collection rotation, or filter-and-delete strategies.

Design meeting rhythms, metric reporting, quarterly planning, and decision-making velocity for scaling companies.

Optimizes Qdrant query performance for large limits across multiple shards by using Poisson-distributed subsampling to reduce inter-shard data transfer.

Guides scaling decisions for Qdrant vector databases based on data volume, query throughput, latency, or query volume.

Design or revise a company's commercial policy governing discounts, approver thresholds, exception flows, and deal frameworks using data-backed matrices and linting tools.

Tests approaches at extreme scales (1000x bigger/smaller, instant/year-long) to expose fundamental truths hidden at normal scales.

Design and enforce microsegmentation policies using workload identity and label-based rules to prevent lateral movement in zero trust architectures, with guidance for tools like VMware NSX, Illumio, and Calico.

Creates modular typography scales with size, weight, and line-height relationships for consistent digital interfaces.

Design and implement network segmentation using firewall security zones, VLANs, ACLs, and microsegmentation policies to restrict lateral movement and enforce least-privilege network access.

Tests approaches at extreme scales to reveal fundamental limits and strengths hidden at normal scales.

Risk managers should SIZE trades, not BLOCK them. Trigger when: (1) capital manager blocks trades, (2) risk manager rejects orders, (3) trades fail due to 'exceeds limit', (4) components don't work in harmony.

Scaling Vision with Sparse Mixture of Experts

LoRA: Low-Rank Adaptation of Large Language Models

Audits a codebase for scalability bottlenecks, flagging N+1 queries, missing indexes, unbounded result sets, blocking operations, and memory pressure, then outputs a prioritized fix list.

Scan installed skills to extract cross-cutting principles and distill them into rules files.

Foundational Scala patterns covering immutability, pattern matching, traits, case classes, for-comprehensions, and functional programming. Use when writing Scala code, understanding the type system, or needing guidance on which specialized Scala skill to use. This is the entry point for Scala development.

Language-specific super-code guidelines for scala.

Improves AI skills for activation, clarity, and cross-model reliability through benchmarking, salience tuning, and regression triage.

Scaling Vision Transformers to 22 Billion Parameters

Reduce decision time and cognitive burden by limiting the number and complexity of simultaneous choices in user interfaces.

Audit App Store and Google Play listings for discoverability and conversion — keywords, localized metadata, screenshots, ratings prompts — then emit a prioritized ASO plan. Use when "optimize our store listing", "improve app downloads", or "ASO". Submission mechanics → plan-mobile-readiness. Plan only.

Generates four improved variations of any installed skill, scores them against a weighted rubric, and applies the winning version while preserving the original.

Use to design or audit a QA dispute and appeal workflow with timeboxes, adjudication standards, and second-level consistency so appeals improve trust instead of rewriting scores without rules. Trigger for "QA appeal process", "agents disputing scores", "who adjudicates QA disputes", "overturn rate too high", second-level review standards, or calibration erosion from ad-hoc score changes.

Apply Miller's Law and chunking strategies to group information into meaningful units, making interfaces easier to scan, understand, and recall.

AutoAugment: Learning Augmentation Strategies from Data

Scales reinforcement learning with PufferLib — high-throughput parallel training (PuffeRL), vectorized environments, and native multi-agent systems achieving 2-10x speedups over standard implementations. Use when scaling RL to millions of steps per second, running vectorized or multi-agent setups, building custom PufferEnv tasks, or integrating game environments (Atari, Procgen, NetHack, PettingZoo). For standard single-agent algorithm implementations (PPO/SAC/DQN) or quick prototyping prefer alterlab-stable-baselines3. Part of the AlterLab Academic Skills suite.

Provides React composition patterns to avoid boolean prop proliferation, covering compound components, state management, and React 19 API changes.