SuperCompress: Efficient LLM Context Compression

Citation

If you use SuperCompress in your research or project, please cite it as follows:

BibTeX citation for LaTeX documents — copy the block above into your .bib file.

Model Architecture

SuperCompress uses a query-aware neural learned policy that operates on token-level features:

• Input features: Token position, TF-IDF relevance score, named entity density, part-of-speech distribution, line length, and semantic similarity to query
• Policy: Lightweight logistic regression with feature interactions, trained via reinforcement learning to maximize recall under a token budget
• Output: Per-line keep/drop decision with a budget constraint that adapts to context length

Benchmark Results

All benchmarks conducted on CPU (Apple M1) at a fixed 35% token budget on 8 project seeds. Results may vary by hardware and context size.

For more detailed results including adaptive mode and policy comparisons, see the full benchmarks page.

Policy Comparison at 35% Budget

* Summarization latency excludes the extra LLM call cost. SuperCompress requires zero GPU time.

Resources

• GitHub Repository
• PyPI Package — pip install supercompress
• Interactive Comparison Tool
• Full Benchmarks
• Token Compression Guide
• Playground
• Awesome Token Compression List