Gnonymous/EVOKE-ALFWorld-7B
Gnonymous/EVOKE-ALFWorld-7B is a 7.6 billion parameter language model based on Qwen2.5-7B-Instruct, specifically enhanced with the EVOKE post-training method. This model is designed to improve goal-directed decision-making by eliciting world knowledge, making it particularly effective for tasks within simulated environments like ALFWorld. It specializes in complex reasoning and planning, leveraging its 32768 token context length for advanced problem-solving.
Loading preview...
EVOKE-ALFWorld-7B: Enhanced for Goal-Directed Decisions
Gnonymous/EVOKE-ALFWorld-7B is a 7.6 billion parameter model built upon the Qwen2.5-7B-Instruct architecture. Its core innovation lies in the EVOKE post-training method, which is specifically engineered to enhance the model's ability to leverage world knowledge for making goal-directed decisions. This makes it particularly adept at navigating and solving problems within interactive, simulated environments.
Key Capabilities
- World Knowledge Elicitation: Utilizes the EVOKE method to better access and apply relevant world knowledge during decision-making processes.
- Goal-Directed Decision Making: Optimized for tasks requiring sequential actions and strategic planning to achieve a specific objective.
- ALFWorld Performance: Specifically fine-tuned and evaluated for performance in the ALFWorld environment, demonstrating its proficiency in complex interactive tasks.
- Large Context Window: Benefits from a 32768 token context length, allowing it to process and understand extensive task descriptions and histories.
When to Use This Model
This model is ideal for research and applications focused on:
- Embodied AI and Agents: Developing AI agents that need to interact with and make decisions in virtual environments.
- Complex Reasoning Tasks: Scenarios where understanding and applying world knowledge is crucial for problem-solving.
- Simulated Environments: Tasks similar to ALFWorld that require planning, execution, and adaptation based on environmental feedback.
For more technical details, including the underlying research and evaluation scripts, refer to the project's GitHub repository and the associated paper.