WeWorkBuddy

首页 / 浏览

全部内容

浏览全部已收录的公开内容,包括尚未归入主分类的条目。

18978 条公开内容

arXiv ·

Quantiles, Not Points: Streaming Quantile Matching for Ordinal-Aware LLM-as-a-Judge Calibration

arXiv:2611.00447v1 Announce Type: new Abstract: LLM-as-a-judge scores are often distributionally shifted from human ratings on any given evaluation task, and post-hoc calibration (mapping LLM to human scores in a small labeled anchor set) has become a popular method to correct for this shift. In this paper we derive the calibration from two positions: that LLMaJ scores are ordinal ratings drawn from a distribution rather than isolated points. We present streaming quantile matching (SQM), a versi

arXiv ·

From Small to Large: Assessing Medical Reasoning, Benchmark Contamination, and Model Size in Language Models

arXiv:2611.00460v1 Announce Type: new Abstract: Large language models (LLMs) now answer medical exam questions at or above expert level, but it is unclear whether such scores reflect clinical competence, or are instead affected by the format of the questions being asked (e.g., multiple-choice versus open-ended), prior exposure to the benchmarks, and the difficulty of the questions. These benchmarks are also public and widely used online, which may be resulting in LLM memorization. Models may per

arXiv ·

Exo: An Architecture for Self-Improving Systems at Runtime

arXiv:2611.00463v1 Announce Type: new Abstract: Agent harnesses govern how language models act in an environment; designing their policies by hand is difficult, while automated techniques require representative datasets and costly executions. We present Exo, an agent runtime architecture for safe learning during ongoing work that collapses the optimization loop into task execution, allowing agents to revise both their harness and their strategies for further improvement. A minimal protected mech

arXiv ·

Compression Technique Interactions in Small Language Models: A Factorial Ablation Study

arXiv:2611.00464v1 Announce Type: new Abstract: Deploying small language models on resource-constrained hardware requires compression, yet practitioners lack systematic guidance on how quantization, pruning, and post-compression recovery interact when combined. We present a factorial ablation study evaluating 72 compression configurations across four SLM architectures (2.7B-7B parameters), extended with a 48-configuration ordering analysis, benchmark accuracy validation on 30 compressed models,

arXiv ·

Few Trials, Many Metrics: Beta Mixtures for Low-Budget AI Evaluation

arXiv:2611.00474v1 Announce Type: new Abstract: Evaluating a stochastic AI system often requires many repeated samples per instance, especially for metrics defined over long horizons. Recent work has shown that modeling instance-level success probabilities with a Beta distribution recovers Pass@$k$ from far fewer samples than direct estimation. We show that this framework supports a broader range of evaluation targets. We fit a mixture of Beta distributions by marginal maximum likelihood, with t

arXiv ·

Swap, Replicate, Read the Mass: A Protocol for Auditing Demographic Bias in System One Decision Models

arXiv:2611.00496v1 Announce Type: new Abstract: System One decision models replace text generation with typed, probability-scored answers and are entering production. Answers execute without a person in the loop, and no multi-model audit exists on established bias benchmarks. We provide that audit: three benchmarks: BBQ (all eleven categories), CrowS-Pairs, and SilverSpoon, adapted to the typed interface, twelve models plus a no-thinking ablation, and 1,638,000 decisions. Every order-sensitive i

arXiv ·

Robustness Starts with the Objective: Designing and Certifying Robust LLM-Generated Heuristics

arXiv:2611.00504v1 Announce Type: new Abstract: LLMs are increasingly used to design heuristics for computer systems. However, most synthesis frameworks evaluate candidates by their average performance, which can mask severe regressions on specific workloads. Even adding adversarial training inputs doesn't eliminate these regressions, and percentile objectives (e.g., optimizing for the 95th percentile) overlook how severe regressions can be beyond the selected threshold. We present TAILOR, a fra

Product Hunt ·

Tao Browser

Private browser with a separate login for every company

Product Hunt ·

Brink

Turn your MacBook trackpad's edges into controls

GitHub Discovery ·

govijr/terminal-mac

Entire MacOS running in your terminal.

GitHub Discovery ·

CzJam/gjgjxweb

《搞机工具箱》Web移植版

Product Hunt ·

One Roof

A new way to build with agents. You stay in charge.

V2EX ·

[推广] [中转站推广] 双 11 活动新用户注册送 10 元

站点: https://onetoken.uk/ 147 个模型,你想要的都有 不降智分组:GPT Pro 满血高智商、gpt-astra 不降智、Codx Pro DeepSeek 走官方定价 0.2x (峰谷价) 生图 1-4k 一张 $0.04 本站特色福利分组,opus4.6 满血,0.04x 留下用户名领鸡蛋。

GitHub Discovery ·

shanselman/gputaskman

A GPU task manager for Windows, Linux, and WSL: per-process VRAM, GPU activity, and LLM model/KV-cache insights for NVIDIA, AMD, and Intel GPUs (plus Huawei Ascend on Linux).

Hacker News ·

Show HN: Project Wisteria-A BYOK chatbot that saves thoughts and memory locally

Its name is 'Wisteria'. A conversation focused desktop chat bot using on BYOK. I built it to address weaknesses about LLM's long-term memorys through a memory architecture of my own. The idea: Arrived at was a memory to compression mechanism that through several level of abstraction. Usually, abstract memory compression is increase the hallucinations. But i tried to address this with an my own architecture. I new it later: some mecahnism to same name now, But last year i called it 'Anchoring'. I

V2EX ·

[分享创造] 这可能是今年最让 JS 开发者激动的事情了

上面是我用 JS+HTML+WebGL2 做的鹈鹕 3D 骑自行车,FPS 也非常不错。 你可能会说“这不是用 Ai 有手就行了么?” 但是,如果我说我把这上面用 xbintsc 编译成原生二进制,还不到 10MB,还能在没有 WebView 环境上跑,还能支持 Node.js 和 HTML 混写,如下: <html> <body> <div id="out">loading...</div> <script lang="ts"> import { readFileSync } from "fs"; const out = document.getElementById("out"); const world = readFileSync("data.txt", "utf8"); out.textContent = "hello " + world; </script> </body> </html> 要知道以前光 chrome 最少也得有 100MB ,用 xbintsc 编译产物 10M

V2EX ·

[推广] Codex 自建池|高性能兼顾性价比

• 纯 Pro 池,多线路稳定供应,成功率长期 99.9+ • Codex 全模型可用 • 高 RPM / 高并发,适合长时间持续跑 • 按量全场 0.2 ,充值立享 0.17 优惠,月度阶梯最低 0.15 • 包月订阅多档可选,最低直达 0.15 • 订阅套餐还可以共享官方重置 • 源头站,可接下游,注册送免费额度( agent 延迟发放) 地址: codex.zhuanyedaili.com ⁠ 联系方式:QQ 群:1106944249 https://codex.zhuanyedaili.com/scripts/codex-qq-group.jpg

V2EX ·

[推广] [LabEx] 面向 Linux , DevOps,网络安全初学者的双十一特惠活动

LabEx 是面向 Linux ,DevOps ,网络安全初学者的实践教学平台,我们针对中国大陆地区的用户提供了专属的双十一折扣, 与黑色星期五同价,提前享受黑五促销价。这是全年唯一的一次集中促销活动,也是折扣力度最大的一次。 如果你有兴趣入门 Linux ,DevOps ,Cybersecurity ,那么 LabEx 非常值得一试,我们提供了免费计划便于你充分尝试和体验,有兴趣的话可以趁着双十一活动升级。 LabEx 针对国内用户提供了「支付宝」付款方式,如果你在海外上学或工作,也可以使用你偏好的其他付款方式。 LabEx 网站: https://labex.io/learn 了解我们的设计理念: https://labex.wiki 直达促销页面: https://labex.io/zh/pricing