https://paperpulse-d8w.pages.dev en-us Daily AI research digest — top arXiv papers explained in plain English by Alex and Sam. PaperPulse AI v1.0 © 2026 PaperPulse AI PaperPulse AI PaperPulse AI no episodic Daily AI research digest — top arXiv papers explained in plain English. PaperPulse AI podcast@example.com https://paperpulse-d8w.pages.dev/cover.png https://paperpulse-d8w.pages.dev https://paperpulse-d8w.pages.dev/episodes/paperpulse_20260917_184205.mp3 Today's episode covers 2 papers: Objective vs. Search: Decomposing What Makes a Good Tokeniser; and A Zeroth-Order Paradigm for LLM Preference Alignment. Thu, 17 Sep 2026 18:42:59 +0000 paperpulse-paperpulse_20260917_184205.mp3 #5: Objective vs. Search: Decomposing What M + A Zeroth-Order Paradigm for LLM Preferen Today's episode covers 2 papers: Objective vs. Search: Decomposing What Makes a Good Tokeniser; and A Zeroth-Order Paradigm for LLM Preference Alignment. 4:52 no full 5 <p>Today's episode covers 2 papers: Objective vs. Search: Decomposing What Makes a Good Tokeniser; and A Zeroth-Order Paradigm for LLM Preference Alignment.</p> <p><strong>Papers covered:</strong></p> <ul><li><a href="https://arxiv.org/abs/2609.19145v1">Objective vs. Search: Decomposing What Makes a Good Tokeniser</a></li> <li><a href="https://arxiv.org/abs/2609.19144v1">A Zeroth-Order Paradigm for LLM Preference Alignment</a></li></ul> https://paperpulse-d8w.pages.dev/episodes/paperpulse_20260917_181807.mp3 Today's episode covers 2 papers: Objective vs. Search: Decomposing What Makes a Good Tokeniser; and A Zeroth-Order Paradigm for LLM Preference Alignment. Thu, 17 Sep 2026 18:20:16 +0000 paperpulse-paperpulse_20260917_181807.mp3 #4: Objective vs. Search: Decomposing What M + A Zeroth-Order Paradigm for LLM Preferen Today's episode covers 2 papers: Objective vs. Search: Decomposing What Makes a Good Tokeniser; and A Zeroth-Order Paradigm for LLM Preference Alignment. 5:16 no full 4 <p>Today's episode covers 2 papers: Objective vs. Search: Decomposing What Makes a Good Tokeniser; and A Zeroth-Order Paradigm for LLM Preference Alignment.</p> <p><strong>Papers covered:</strong></p> <ul><li><a href="https://arxiv.org/abs/2609.19145v1">Objective vs. Search: Decomposing What Makes a Good Tokeniser</a></li> <li><a href="https://arxiv.org/abs/2609.19144v1">A Zeroth-Order Paradigm for LLM Preference Alignment</a></li></ul> https://paperpulse-d8w.pages.dev/episodes/paperpulse_20260917_181224.mp3 Today's episode covers 2 papers: Objective vs. Search: Decomposing What Makes a Good Tokeniser; and A Zeroth-Order Paradigm for LLM Preference Alignment. Thu, 17 Sep 2026 18:13:17 +0000 paperpulse-paperpulse_20260917_181224.mp3 #3: Objective vs. Search: Decomposing What M + A Zeroth-Order Paradigm for LLM Preferen Today's episode covers 2 papers: Objective vs. Search: Decomposing What Makes a Good Tokeniser; and A Zeroth-Order Paradigm for LLM Preference Alignment. 3:46 no full 3 <p>Today's episode covers 2 papers: Objective vs. Search: Decomposing What Makes a Good Tokeniser; and A Zeroth-Order Paradigm for LLM Preference Alignment.</p> <p><strong>Papers covered:</strong></p> <ul><li><a href="https://arxiv.org/abs/2609.19145v1">Objective vs. Search: Decomposing What Makes a Good Tokeniser</a></li> <li><a href="https://arxiv.org/abs/2609.19144v1">A Zeroth-Order Paradigm for LLM Preference Alignment</a></li></ul> https://paperpulse-d8w.pages.dev/episodes/paperpulse_20260917_180611.mp3 Today's episode covers 2 papers: Objective vs. Search: Decomposing What Makes a Good Tokeniser; and A Zeroth-Order Paradigm for LLM Preference Alignment. Thu, 17 Sep 2026 18:06:42 +0000 paperpulse-paperpulse_20260917_180611.mp3 #2: Objective vs. Search: Decomposing What M + A Zeroth-Order Paradigm for LLM Preferen Today's episode covers 2 papers: Objective vs. Search: Decomposing What Makes a Good Tokeniser; and A Zeroth-Order Paradigm for LLM Preference Alignment. 7:21 no full 2 <p>Today's episode covers 2 papers: Objective vs. Search: Decomposing What Makes a Good Tokeniser; and A Zeroth-Order Paradigm for LLM Preference Alignment.</p> <p><strong>Papers covered:</strong></p> <ul><li><a href="https://arxiv.org/abs/2609.19145v1">Objective vs. Search: Decomposing What Makes a Good Tokeniser</a></li> <li><a href="https://arxiv.org/abs/2609.19144v1">A Zeroth-Order Paradigm for LLM Preference Alignment</a></li></ul> https://paperpulse-d8w.pages.dev/episodes/paperpulse_20260917_175526.mp3 Today we explore tokeniser design decomposition and zeroth-order preference alignment without gradients. Thu, 17 Sep 2026 18:02:10 +0000 paperpulse-paperpulse_20260917_175526.mp3 #1: Objective vs. Search: Decomposing What M + A Zeroth-Order Paradigm for LLM Preferen Today we explore tokeniser design decomposition and zeroth-order preference alignment without gradients. 4:05 no full 1 <p>Today we explore tokeniser design decomposition and zeroth-order preference alignment without gradients.</p> <p><strong>Papers covered:</strong></p> <ul><li><a href="https://arxiv.org/abs/2609.19145v1">Objective vs. Search: Decomposing What Makes a Good Tokeniser</a></li> <li><a href="https://arxiv.org/abs/2609.19144v1">A Zeroth-Order Paradigm for LLM Preference Alignment</a></li></ul> Thu, 17 Sep 2026 18:42:59 +0000