learn.ifs.name

DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning

DeepSeek-AI, 2025, DeepSeek-R1 (arXiv preprint arXiv:2501.12948)

Subscriber deck — free during early access

1 cards in “5 Conclusion, Limitations, and Future Work” · whole deck

5 Conclusion, Limitations, and Future Work

  1. What limitations of DeepSeek-R1 does the report list?
    It is weaker than DeepSeek-V3 at function calling, multi-turn dialogue, role-play and JSON output; it may reason or answer in English (or Chinese) for queries in other languages; few-shot prompting hurts it, so zero-shot prompts are recommended; and it gained little on software engineering, where slow evaluation limited RL.
    • DeepSeek-AI, 2025, DeepSeek-R1 (arXiv preprint arXiv:2501.12948), §5

Source

  • DeepSeek-AI. 2025. DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning. In arXiv preprint arXiv:2501.12948. https://arxiv.org/abs/2501.12948