#KnowledgeRetrieval

2024-12-08

Reverse engineering LlamaIndex has been a fascinating dive into understanding how its Workflow processes are structured to handle dynamic data retrieval and integration seamlessly. At its core, a LlamaIndex Workflow orchestrates the interaction between indexes, queries, and retrieval logic, ensuring efficient and context-aware results. By analyzing its modular design, I found that each task—whether building an index or querying it—is highly decoupled, enabling scalability and customization. The workflow’s use of adaptive heuristics and stateful operations allows it to fine-tune results in real-time while handling diverse data sources. This design not only ensures flexibility but also showcases how workflows in LlamaIndex intelligently manage complexity in knowledge retrieval tasks. Understanding these processes provides valuable insights for building robust, modular AI systems. #LlamaIndex #ReverseEngineering #AIWorkflows #KnowledgeRetrieval

2024-01-04

Interesting survey paper that summarizes 32 methods to mitigate hallucination in LLMs.
S.M Towhidul Islam Tonmoy et al,
A Comprehensive Survey of Hallucination Mitigation Techniques in Large Language Models
arxiv.org/abs/2401.01313

#llms #hallucination #rag #knowledgeretrieval #cove

Taxonomy of hallucination mitigation techniques in LLMs, focusing on prevalent methods that involve model
development and prompting techniques. Model development branches into various approaches, including new decoding
strategies, knowledge graph-based optimizations, the addition of novel loss function components, and supervised fine-tuning.
Meanwhile, prompt engineering can involve retrieval augmentation-based methods, feedback-based strategies, or prompt tuning.

Client Info

Server: https://mastodon.social
Version: 2025.04
Repository: https://github.com/cyevgeniy/lmst