v0.4: control-theoretic rewrite, fixed loop detector, ablations, PCA, perplexity tracking

Add Colab benchmark + chat script for Qwen3-8B-AWQ

v0.3: comprehensive README with SOTA comparison, benchmarks, and new features

Add v0.3 GPT-2 benchmark results

v0.3: update setup.py version

v0.3: core with save/load calibration + auto-tune

v0.3: correctors unchanged from v0.2

v0.3: detectors with export/load calibration

v0.3: config with auto-tune and profile path

v0.3: add calibration profiles, auto-tune, benchmarks

v0.2 README: document calibration-first design

v0.2 demo: updated for calibration-first design

v0.2 dashboard

v0.2 core: budget-limited corrections, calibration-aware R tracking

v0.2 correctors: scale-normalized, budget-aware

v0.2 detectors: calibration-first with statistical thresholds

v0.2: calibration-first, budget-limited - fixes over-correction

v0.2: calibration-first, budget-limited, no over-correction

Add demo.py

Add setup.py

Add aria_llm/dashboard.py

Add aria_llm/correctors.py

Add aria_llm/detectors.py

Add aria_llm/core.py

Upload aria_llm/config.py

Upload aria_llm/core.py

Upload aria_llm/__init__.py

Add ARIA: Adaptive Reliability & Integrity Attachment for LLMs

initial commit