<![CDATA[llm learning road]]> https://rss.com/podcasts/zcs-llr-learning-road llm learning road

]]> RSS.com 2026.319.134332 Thu, 19 Mar 2026 21:06:04 GMT zh 7edc257a-4918-543e-97f3-1b774a770591 https://media.rss.com/zcs-llr-learning-road/20260317_060320_e050fdb8ef91d2b0e0641d31c6cbad50.png llm learning road https://rss.com/podcasts/zcs-llr-learning-road yes zc zc false episodic podcast Seattle, Seattle, King County, Washington, 98104, USA <![CDATA[chatgpt-pro-LLM进化史]]> chatgpt-pro-LLM进化史

]]> https://rss.com/podcasts/zcs-llr-learning-road/2643997 061ecd47-63ca-4e44-8e92-35b1ca948e1b 4819 full false Thu, 19 Mar 2026 21:06:01 GMT <![CDATA[minimax-LLM发展史与关键技术演进深度研究报告]]> minimax-LLM发展史与关键技术演进深度研究报告

]]> https://rss.com/podcasts/zcs-llr-learning-road/2643762 610b8ec3-23ab-47d2-87cc-f396ceb8dfbf 4263 full false Thu, 19 Mar 2026 20:53:44 GMT <![CDATA[doubao-大模型+LLM+发展脉络梳理]]> doubao-大模型+LLM+发展脉络梳理

]]> https://rss.com/podcasts/zcs-llr-learning-road/2643695 a1133b95-602c-407a-8c13-3237741606e5 5274 full false Thu, 19 Mar 2026 19:01:52 GMT <![CDATA[qwen-大模型进化图谱：从技术跃迁、厂商博弈到开发者生存指南]]> qwen-大模型进化图谱：从技术跃迁、厂商博弈到开发者生存指南

]]> https://rss.com/podcasts/zcs-llr-learning-road/2643661 fbc08faa-396b-40b2-bd4c-176fff4f5625 1887 full false Thu, 19 Mar 2026 18:49:21 GMT <![CDATA[DeepseekMath]]> 00:00:00 Start

00:17:41 DeepSeekMathBase+7B+的数学能力评估

00:27:47 DeepSeekMath-RL+的训练与评估

00:40:24 探索抗噪声奖励信号的算法

00:46:35 DPO、PPO+和+GRPO+的目标及梯度

00:51:30 Closing

]]> https://rss.com/podcasts/zcs-llr-learning-road/2642983 51cabf7a-3b2c-47da-bb78-d47478a72f5c 3090 full false Thu, 19 Mar 2026 16:09:28 GMT <![CDATA[instructGPT]]> 00:00:00 Start

00:18:49 InstructGPT：在不同方面的表现与改进

00:28:34 Closing

]]> https://rss.com/podcasts/zcs-llr-learning-road/2642981 040080a0-9496-4b04-a714-973fe2183b46 1713 full false Thu, 19 Mar 2026 16:08:27 GMT <![CDATA[ReAct]]> 00:00:00 Start

00:19:25 ReAct+Outperforms+Other+Methods+in+Tasks

00:28:34 ReAct+在+ICLR+2023+会议论文中发布

00:34:25 ICLR+2023+会议论文中的实验

00:43:45 ReAct+在+ALFWorld+中的刀具清洁问题

00:51:15 Closing

]]> https://rss.com/podcasts/zcs-llr-learning-road/2642975 dca25eb0-3212-4b38-8afc-b58dde6a3322 3075 full false Thu, 19 Mar 2026 16:07:45 GMT <![CDATA[deepseek V3.2]]> deepseek V3.2

00:00:00 Opening

00:30:00 Main Discussion

]]> https://rss.com/podcasts/zcs-llr-learning-road/2637009 93d02d23-6f15-41e7-86f7-a2d0586e256a 2420 full false Tue, 17 Mar 2026 18:32:49 GMT <![CDATA[deepseek-R1]]> DeepSeek R1 full discussion and segmented playback.

OUTLINE

00:00:00 Opening

00:16:02 DeepSeek R1 - Segment 1

00:29:54 DeepSeek R1 - Segment 2

00:47:34 DeepSeek R1 - Segment 3

00:56:22 DeepSeek R1 - Segment 4

01:04:21 DeepSeek R1 - Segment 5

01:15:42 DeepSeek R1 - Segment 6

01:26:49 DeepSeek R1 - Segment 7

01:42:33 DeepSeek R1 - Segment 8

02:00:42 DeepSeek R1 - Segment 9

02:12:45 DeepSeek R1 - Segment 10

02:21:42 DeepSeek R1 - Segment 11

02:24:38 Ending

]]> https://rss.com/podcasts/zcs-llr-learning-road/2637005 97324d1c-4fae-440a-8aae-76bbb2d5648e 9009 full false Tue, 17 Mar 2026 18:32:48 GMT <![CDATA[硅谷坐标 x 田渊栋： 解析大模型护城河、记忆存储瓶颈与Agent对社会冲击]]> 田渊栋（Meta AI/FAIR，强化学习与大模型推理/长上下文方向）分享对大模型竞争格局、模型“记忆”机制、长上下文与存储瓶颈、推理路线，以及 Agent 时代社会冲击的判断与观察。

### 内容要点

- 大模型护城河：数据/算法/infra/人才

- 两种记忆：上下文（工作记忆）与权重（世界观）

- 长上下文与存储供应链瓶颈

- Agent 形态与社会影响

]]> https://rss.com/podcasts/zcs-llr-learning-road/2635528 5bf0eb42-d075-4005-b8e5-9f4d637c8cad 3709 full false Tue, 17 Mar 2026 07:03:35 GMT