Write-up implements temperature, top-p, min-p and repetition penalty in code
11/19/2026 — 11/20, 14:19·1 sources (all)·1 reports (all)
Key takeaways
- A technical write-up implements the main LLM sampling strategies — temperature, top-p, min-p and repetition penalty — in code and walks thro…
Background & analysis
On November 19, 2026 at 21:25, a post titled "Sampling Strategy Implementation: temperature, top-p, min-p, and Repetition Penalty" appeared on the Juejin platform, focusing on the sampling strategies used in LLM inference. The article implements the main sampling strategies in code, covering temperature, top-p, min-p, and repetition penalty, and walks through the failure mode of each one individually.
The author argues that these sampling parameters are not parallel options to be chosen from, but a group of components with clearly defined semantic roles. Under the same model, swapping in a different set of sampling parameters can noticeably change the style of the output. The article characterizes sampling strategies as the components with the fewest parameters yet the most direct impact in an inference system: for a single model, moving from restrained to erratic output can come down to a single line of code, since the difference lies only in the sampling configuration.
So far, only this one article has been reported on the topic. The matter currently rests at the point of the article being published, with no further developments or follow-up coverage reported.
AI-generated from 1 reports · updated 2 days ago
- Heat index
- 25
- All sources
- 1
- All reports
- 1
- First seen
- 3 days ago
Reports on this story headlines open the original
The article implements the main LLM sampling strategies in code—temperature, top-p, min-p, and repetition penalty—and walks through how each one fails. Rather than interchangeable knobs, the author argues they are components with distinct semantic roles, and swapping them can change a model's output style dramatically.
掘金AI score 55
Other stories people are talking about
- Heat index 1143RisingAnthropic Reaches $1.6 Trillion Market Value in First Trading Week After IPO21 sources
- Heat index 1011Anthropic Mathematician Says Claude Formalized the Generalized Riemann Hypothesis21 sources
- Heat index 622Anthropic IPO total raise reaches $93.5 billion after greenshoe exercise20 sources
- Heat index 596RisingThinking Machines Lab Open-Sources Inkling 2 Mini with 32GB Model Card Recommendation10 sources
- Heat index 490RisingSEC probes Nvidia accounting practices as shares post seven-week slide9 sources
How is heat calculated?About the methodHide
Heat counts how many independent sources covered a story in the last 48 hours: one source counts once no matter how many posts it published, decaying with a 24-hour half-life. What ranks first is what many people are talking about.
This page aggregates public feeds. Headlines and summaries are machine-organized and remain the property of the original authors; verify important facts at the source. If you believe a headline or summary infringes your rights, email the contact address in the footer with the page URL and basis of your claim, and we will remove or replace it after review.
- Surge
- Discussion rising fast
- New
- First report within 6 hours
- Rising
- Still gathering discussion

