long-context

Malicious tool definition detected

Tool: SKILL.md [1/2] Description: --- name: long-context description: Extend context windows of transformer models using RoPE, YaRN, ALiBi, and position interpolation techniques.

Tool: SKILL.md [2/2] Description: | **Position Interpolation** | 32k+ | Minimal (1k steps) | Moderate | Poor (by design) | Quick extension | ## Implementation Patterns ### HuggingFace Transformers Integration ```python from transformers import AutoModelForCausalLM, AutoConfig # RoPE with YaRN scaling config = AutoConfig.from_pretrained("mistralai/Mistral-7B-v0.1") config.rope_scaling = { "type": "yarn", "factor": 8.0, "original_max_position_embeddings": 8192, "attention_factor": 1.0 } model = Au

Malicious tool definition detected

Tool: references/extension_methods.md [1/2] Description: # Context Extension Methods Comprehensive comparison of YaRN, ALiBi, and Position Interpolation based on published research.

Tool: references/extension_methods.md [2/2] Description: Head 8: Weak recency bias (slope=0.0625) ``` ### Advantages 1.

Malicious tool definition detected

Tool: references/fine_tuning.md [1/2] Description: # Fine-tuning for Context Extension Complete guide to fine-tuning transformer models for longer context windows.

Tool: references/fine_tuning.md [2/2] Description: lazy dog.

Malicious tool definition detected

Tool: references/rope.md [1/2] Description: # RoPE: Rotary Position Embeddings Complete technical guide based on RoFormer paper (arXiv 2104.09864) and HuggingFace transformers implementation.

Tool: references/rope.md [2/2]