Recursa Paper Shows RSI Lifts Model Performance Without Fine-Tuning
The non-obvious takeaway is that domain adaptation can be achieved through a validated skill repository rather than weight updates, making improvement cycles cheaper and reversible. Teams adapting agents to specialised domains such as law can try the repository approach before investing in fine-tuning or expert annotation. The code and paper are on arXiv, with a companion walkthrough at academy.dair.ai.
A new paper proposes Recursa, a recursive self-improvement system for foundation models. Recursa raised DeepSeek-V4.1-Flash from 30.0% to 55.0% all-pass on held-out legal tasks, without a stronger teacher or human experts. The gains come from a self-written knowledge repository storing domain skills, usage instructions, memories and reflections, with parallel learners proposing fixes that are merged only after validation. On Harvey LAB's Firm Knowledge benchmark, Qwen3.8-Flash improved from 35.0% to 46.0% using the same loop, while model weights stayed unchanged. The paper is available at arxiv.org/abs/2611.04035.
Banger paper on RSI with foundation model systems. (bookmark it) Seriously, I highly recommend this paper if you want to start experimenting with RSI in your own company. Lots of great insights and tips. They propose Recursa, which raises DeepSeek-V4.1-Flash from 30.0% to 55.0% all-pass on held-out legal tasks, without a stronger teacher or human experts. The model weights stay the same. The gain comes from a knowledge repository the system writes for itself. The repository stores domain skills, instructions on when to use them, memories of past tasks, and reflections. Recursa first builds it from domain materials. Then parallel learners run tasks and propose fixes, and only validated fixes get merged back. On Harvey LAB's Firm Knowledge benchmark, Qwen3.8-Flash goes from 35.0% to 46.0% with the same loop. If you are adapting an agent to a new domain, try a validated skill repository before reaching for fine-tuning. arxiv.org/abs/2611.04035 Chat with Paper: academy.dair.ai/papers/recursa… 💬 17 🔄 1 ❤️ 28 👀 2645 📊 21 ⚡
