AlexWortega/openjev

A fine-tuned Qwen3.5-4B checkpoint provides a cross-encoder for natural language inference, supporting entailment, contradiction, and neutral classification. The repository documents usage for reranking and game-state evaluation via a custom head. While the model is loadable and usable, it is a specialized fine-tune of an existing base model rather than a new foundation release or major architectural innovation.