Context, examined in public.
Technical explorations of graph centrality, token economics, attention saliency, and provenance-aware compression.
Featured article · 20 min read
LLMSlim vs LLMLingua-2 vs LongLLMLingua: A Reproducible Prompt Compression Benchmark
One reader, one token budget, 300 paired cases on NaturalQuestions, and every script and per-case result published. We compared LLMSlim 0.8 with LLMLingua-2 and LongLLMLingua on 20-document NaturalQuestions at a 25% token budget, with Claude Haiku 5.5 as the reader. With its optional 22M-parameter reranker, LLMSlim was level with LongLLMLingua (78.7 vs 78.0) and 9.7 points ahead of LLMLingua-2, with a median compression time per case of 0.09 s against 9.3 s and 2.3 s.
Every comparison is paired: both systems answer the same cases, so each case gives a score difference of −1, 0 or +1. The 95% interval resamples those differences with replacement 4,000 times, by case or by whole question, and reads the 2.5th and 97.5th percentiles of the resampled means.
Algorithms
2 articlesGraph Centrality & TF-IDF Vectorization for In-Context Redundancy Reduction
Mathematical Derivation of LexRank Stationary Distributions and Priority Tier Filtering
Quadratic Attention Scaling O(N^2) & In-Context Token Reduction Economics
Deriving Computation Savings in Transformer Self-Attention Layers
Systems & Architecture
3 articlesDeterministic Safety Shields: Priority Tier Protection Mechanisms
Formal Syntactic Safeguards for Directives, Code Blocks, and Structured Data
AST Syntax-Aware Normalization for JSON, XML, and YAML Prompts
Compressing Structural Data Payloads Without Invocation Failures
Reverse Proxy Gateway Integration: Context Compression in Production Python Gateways
Deploying Transparent Pre-Dispatch Context Optimization in Enterprise Services
Context Engineering
1 articleMitigating U-Shaped Attention Recall Decay in Long Context Prompts
Structuring Information Density to Overcome Position-Dependent Attention Loss
Benchmarking Methodology
2 articlesLLMSlim vs LLMLingua-2 vs LongLLMLingua: A Reproducible Prompt Compression Benchmark
One reader, one token budget, 300 paired cases on NaturalQuestions, and every script and per-case result published
Architectural Comparison: Offline Graph Methods vs. Neural Perplexity Pruning
Evaluating Throughput, Hardware Footprint, and Latency Metrics for Edge & Cloud Gateways