Ideas & society2024 · shared on Hacker News
Training LLMs to Reason in a Continuous Latent Space
Explore Training LLMs to Reason in a Continuous Latent Space, its original source and Hacker News discussions. Discover community perspectives and save it to your reading list.
GO BEYOND THE ABSTRACT
The conversation around the idea.
Original Hacker News threads. Different perspectives, questions, and things worth thinking about.
Showing up to 100 related threads. A paper may appear in a comment rather than the original submission.
Links found in the discussion
Up to 50 recent source links, in HN item order.
Comment #49328313 Comment #48632178 Comment #47798505 Comment #47602340 Comment #46362877 Comment #44586702 Comment #44586541 Comment #44310690 Comment #44075124 Comment #43599964 Comment #43082648 Comment #42919898 Comment #42831667 Comment #42564877 Comment #42557595 Comment #42556535 Comment #42489805 Comment #42479898 Comment #42385644
Ymaurycy ·
OpenAI O3 breakthrough high score on ARC-AGI-PUB
1,724 points1,755 comments
Ygradus_ad ·
DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via RL
1,351 points1,056 comments
Ygeorgehill ·
The Llama 4 herd
1,235 points658 comments
Yowenthejumper ·
GPT-5 is behind schedule
582 points1,097 comments
YTaurenHunter ·
Coconut by Meta AI – Better LLM Reasoning with Chain of Continuous Thought?
362 points156 comments
Yomarsar ·
Training LLMs to Reason in a Continuous Latent Space
283 points114 comments
Yhiguidebot ·
Watch R1 "think" with animated chains of thought
267 points75 comments
Yfofoz ·
Efficient Reasoning with Hidden Thinking
172 points43 comments
Ynyrikki ·
Beyond Semantics: Unreasonable Effectiveness of Reasonless Intermediate Tokens
138 points66 comments
Ymfiguiere ·
Chain of thought monitorability: A new and fragile opportunity for AI safety
134 points64 comments
Ymarojejian ·
Universal Reasoning Model (53.8% pass 1 ARC1 and 16.0% ARC 2)
122 points28 comments
Ymeetpateltech ·
Claude Opus 4.7
81 points93 comments
Ydanielmorozoff ·
Reasoning by Superposition: A Perspective on Chain of Continuous Thought
60 points1 comments
Y0o_MrPatrick_o0 ·
Claude Code's "extended thinking" is a summary- not authentic thinking
21 points12 comments
Ybilsbie ·
Qwen 3.8 27B is excellent, but it defaults to overthinking things
5 points1 comments
Yt0lo ·
Coconut from Meta lets language models think without words
2 points2 comments
Ysorenjan ·
TinyLoRA – Learning to Reason in 13 Parameters
1 points0 comments
Ythunderbong ·
Training Large Language Models to Reason in a Continuous Latent Space [pdf]
1 points0 comments