LLMSlim Uses Deterministic Pipeline for Prompt Compression Edge Cases

LLMSlim is a Python library that compresses LLM prompts via a six-stage deterministic pipeline, including TF-IDF vectorization and LexRank scoring, without using neural embeddings. It handles edge cases like broken code fences and truncated schemas. The library aims to reduce token counts while preserving critical instructions.