概述
In its standard formulation tau specifies a target cross-entropy in bits, corresponding to perplexity 2^tau; it adjusts a truncation threshold, without guaranteeing every passage’s perplexity or factual correctness.
深入探讨
Standard top-k and top-p decoding apply fixed cutoffs, while the predictability of generated text can vary across a passage. Mirostat, proposed by Basu and colleagues in 2020, treats decoding as a feedback-control problem. In the standard formulation, tau is a target average surprisal or cross-entropy in bits; the corresponding perplexity is 2^tau. The sampler measures token surprisal and adjusts a truncation threshold based on the difference from the target. If generated tokens are less surprising than the target, the threshold can be loosened; if they are more surprising, it can be tightened. This aims to regulate average surprise over generation and does not require each token or passage to match the target exactly. It does not measure factual correctness.
现实世界的实施
防止 KoboldAI 等本地法学硕士应用程序中的长故事或角色扮演世代陷入重复循环。
在 llama.cpp 和 Ollama 中作为 mirostat 设置（模式 1 或 2、tau、eta）公开，供爱好者调整输出质量。
稳定聊天机器人的响应，使它们在长时间会话中既不会重复短语，也不会转向不连贯的切线。
供希望在整个生成的段落中保持一致的创造力水平而不是质量波动的作家使用。
Mirostat Perplexity 控制的未来
Mirostat options are exposed in some local inference tools; llama.cpp, for example, documents mode, tau and eta parameters. Availability and behavior depend on the runtime and version. Its control-theoretic framing is inspiring further adaptive decoders that regulate other signals like factuality or diversity. As long-form generation grows, expect feedback-driven sampling to be combined with retrieval and repetition penalties, and possibly auto-tuned tau values that adapt to genre, replacing manual perplexity targets.
付诸实践
为什么这很重要
- 速度与规模. 语言工作流程可以在不牺牲一致性的情况下更快地移动。
- 交通与覆盖范围. 它扩展了跨语言和沟通方式的访问。
- 更清晰的判决. 团队可以花更多时间进行判断，而自动化则可以处理重复。
留意
- 幻觉的事实可以悄悄地进入报告、支持流程或研究成果。
- 及时的敏感性可能会在类似的请求中产生不一致的结果。
- 如果访问控制薄弱，敏感文本数据可能会暴露。
后续步骤
- 在推出之前定义输出格式、语气和质量标准。
- 当准确性很重要时，请使用可信来源进行地面响应。
- 为高风险输出保留人工审查检查点。
- 跟踪故障模式并定期重新训练提示或工作流程。
不断探索
常见问题
What is Mirostat Perplexity Control?
Mirostat is a feedback-based decoding algorithm that aims to regulate the average surprisal of generated tokens toward a target. In its standard formulation tau specifies a target cross-entropy in bits, corresponding to perplexity 2^tau; it adjusts a truncation threshold, without guaranteeing every passage’s perplexity or factual correctness.
What does Mirostat try to control during generation?
Mirostat uses feedback to adjust its sampling threshold toward a target surprise or perplexity level.
Mirostat 与标准 top-k 或 top-p 采样有何根本不同？
Top-k 和 top-p 应用固定阈值，而 Mirostat 根据观察到的意外情况不断调整其截断，就像控制系统一样。
在Mirostat中，参数eta起什么作用？
eta是学习率；它衡量了控制变量 mu 响应观察到的意外值与目标意外值之间的误差而更新的强度。
Mirostat 版本 2 相对于版本 1 的主要改进是什么？
Mirostat 2 消除了对版本 1 中使用的 Zipf 定律假设的依赖，使算法更简单且跨模型更广泛可靠。
Which Mirostat settings does current llama.cpp expose?
llama.cpp documents a Mirostat mode plus tau and eta parameters for its sampling implementation.