evaluation 20
- SageMaker RLVR's Cached Score Does Not Prove Custom-Reward Gains
- Before You Fine-Tune a Jev-Style Model, Write Down the Baseline
- In my rerun, Headroom saved 0 tokens on its GSM8K eval prompts
- mini-AGI's forgetting fix is measured on reading, not on chat
- GameDevBench's 68.8% top score is the only strictly confined row
- Moli's efficiency map blends two runs its own bench keeps apart
- Harness Architecture: Turning Autonomous AI Agent Experiments into Reliable Systems
- AutoRAG-Research: Standardized Benchmarking and Modular RAG Evaluation
- Easy Dataset: Architecture and Pipeline for Synthetic LLM Data and RAG Evaluation
- Prompt Repetition: Approximating Bidirectional Attention in Causal LLMs
- Evaluating Deep Agents CLI on Terminal Bench 2.0 with Harbor
- Source Audit: Anthropic 100k Study and Multi-Agent Productivity Boundaries
- Source Audit: Fara-7B Synthetic Trajectory Engine and Critical-Point Halting
- Source Audit: Andrej Karpathy LLM Council and Deliberation Mechanics
- SIMA 2's Self-Improvement Score Is Not a Game-QA Verdict
- LLM-as-a-Judge Architecture: Enterprise Evaluation with Spring AI Recursive Advisors
- The Production LLM Stack: Architectural Blueprint Across Seven Engineering Layers
- Engineering Production RAG Systems: An Architectural Delivery Roadmap
- CRAG Two Years On: The RAG Benchmark That Scores Silence Above a Wrong Answer
- Building Systematic RAG Workflows: Beyond the Naive Vector Search Pattern