Awesome-LM-SSP

Why we included this project

Awesome-LM-SSP is a manually curated reading list for anyone working on the trustworthiness of large models, with a special focus on multimodal systems like vision-language and diffusion models. It organizes papers and resources into safety, security, and privacy, then breaks those down further into subcategories such as jailbreak, prompt injection, membership inference, model extraction, unlearning, and watermarking. Each entry is tagged by model type (LLM, vision-language, speech, diffusion) and by what it contributes, whether a benchmark, dataset, agent, or defense, so you can filter down to the slice you actually care about. The collection also includes surveys, toolkits, leaderboards, and competitions, which makes it useful both for getting up to speed on a topic and for finding concrete codebases to evaluate. It is maintained through community contributions via issues, so it keeps pace with newly accepted work across major venues.