Enhance benchmark and Cortex modules with new training utilities and improved state management. Update README with example output for Llama-3.2-1B and add training CLI for Cortex module tuning. Refactor scoring functions to reset Cortex state between examples and ensure consistent output. Modify task handling to ensure proper formatting of input data.

Refactor device handling in benchmark and cortex modules to use resolve_torch_device function for improved device selection. Update README with benchmark table formatting and additional device options.

Add benchmark harness documentation to README

Add benchmark harness: run_benchmark.py - CLI entry point

Add benchmark harness: runner.py - Main comparison orchestrator

Add benchmark harness: tasks.py - Standard NLP task definitions

Add benchmark harness: memory_tasks.py - Passkey retrieval and multi-hop memory

Add benchmark harness: __init__.py

Add benchmark harness: scoring.py

Add cortex/adaptive_depth.py

Add cortex/steering_vector.py

Add test suite

Add cortex/memory_bank.py

Add cortex/hallucination_gate.py

Add cortex/pause_and_think.py

Add cortex/backtrack_head.py

Add cortex/__init__.py

Add cortex/core.py — CortexModule, CortexSurgeon, CortexConfig

Add comprehensive README

initial commit