Research map / Large language models / AI and society
RecWizard: A Toolkit for Conversational Recommendation with Modular, Portable Models and Interactive User Interface
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's full text. Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Presents RecWizard, a Hugging Face-based Python toolkit for conversational recommender systems offering modular module/pipeline abstractions, portability via HF hub, LLM integration, and an interactive debugging UI.
Problem
Existing CRS toolkits lack support for pipeline-level evaluation, portability/sharing of models, LLM integration, and interactive user interfaces for debugging and evaluation.
Approach
A toolkit that abstracts CRS into modules (recommender, generator, processor) and pipelines that communicate via natural-language text through HF-compatible tokenizers, with save/push_to_hub portability and a two-mode (INFO/DEBUG) chat interface.
Key results
- RecWizard supports loading pretrained CRS modules from the HF hub with one-liners and building pipelines combining recommender and ChatGPT generator modules.
- Table comparison: RecWizard is the only one of the compared toolkits (CRSLab, FORCE) that is open-source, portable, has a user interface, and is LLM-supported.
Limitations
- No comprehensive unified benchmark evaluation of the included CRS models.
- Models are trained with source-code-like strategies; no standardized training APIs.
- Pipeline-level evaluation is not provided in the toolkit yet.
Experimental setup
Datasets: ReDIAL
Baselines: CRSLab FORCE
Builds on: Hugging Face transformers CRSLab
Closest papers by meaning
- ESPRESSO: An Effective Approach to Passage Retrieval for High-Quality Conversational Recommender Systems (AAAI 2025 · similarity 0.82)
- A Statistical Framework for Ranking LLM-based Chatbots (ICLR 2025 · similarity 0.81)
- RecFlow: An Industrial Full Flow Recommendation Dataset (ICLR 2025 · similarity 0.81)
- Chatbot Arena: An Open Platform for Evaluating LLMs by Human Preference (ICML 2024 · similarity 0.81)
- ConvRec-R1: Training LLM-based Conversational Recommender Systems with Reinforcement Learning (ICLR 2026 · similarity 0.81)
- WildBench: Benchmarking LLMs with Challenging Tasks from Real Users in the Wild (ICLR 2025 · similarity 0.80)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.