mHC: Manifold-Constrained Hyper-Connections · HackerPapers discussions

mHC: Manifold-Constrained Hyper-Connections — mentioned by rvz

Fri, 24 Apr 2026 04:00:08 GMT

The paper is here: [0] Was expecting that the release would be this month [1], since everyone forgot about it and not reading the papers they were releasing and 7 days later here we have it. One of the key points of this model to look at is the optimization that DeepSeek made with the residual design of the neural network architecture of the LLM, which is manifold-constrained hyper-connections (mHC) which is from this paper [2], which makes this possible to efficiently train it, especially with its hybrid attention mechanism designed for this. There was not that much discussion around it some…

Paper and discussion history. A mention is not an endorsement.

mHC: Manifold-Constrained Hyper-Connections — mentioned by vibe42

Fri, 20 Mar 2026 19:25:18 GMT

Interesting to compare with deepseek mHC (Jan 2026), which also changes how residual connections work: https://arxiv.org/abs/2512.24880

Paper and discussion history. A mention is not an endorsement.

mHC: Manifold-Constrained Hyper-Connections — mentioned by yorwba

Mon, 12 Jan 2026 15:55:10 GMT

https://arxiv.org/abs/2512.24880 was published less than two weeks ago, which should explain why it's not more common yet. And it's not that amazing either. It's a slight quality improvement for a slight increase in cost. It's not even clear to me whether it pays for itself.

Paper and discussion history. A mention is not an endorsement.

mHC: Manifold-Constrained Hyper-Connections — HN submission

Fri, 02 Jan 2026 19:11:51 GMT

MHC: Manifold-Constrained Hyper-Connections

1 point; 0 discussion replies on HN. Paper and source history. A submission is not an endorsement.

mHC: Manifold-Constrained Hyper-Connections — HN submission

Fri, 02 Jan 2026 01:23:27 GMT

DeepSeek MHC: Manifold-Constrained Hyper-Connections

4 points; 0 discussion replies on HN. Paper and source history. A submission is not an endorsement.

mHC: Manifold-Constrained Hyper-Connections — HN submission

Thu, 01 Jan 2026 16:31:22 GMT

MHC: Manifold-Constrained Hyper-Connections

12 points; 1 discussion reply on HN. Paper and source history. A submission is not an endorsement.

mHC: Manifold-Constrained Hyper-Connections — mentioned by edflsafoiewq

Thu, 01 Jan 2026 15:35:38 GMT

Link: https://arxiv.org/abs/2512.24880

Paper and discussion history. A mention is not an endorsement.

mHC: Manifold-Constrained Hyper-Connections — mentioned by rvz

Thu, 01 Jan 2026 11:57:09 GMT

Except that the "content" that reaches the top is always about AI / LLMs and nothing else and it is "all the time". Any opportunity to comment, he will link back to his own blog. He even reposted the same link (which is about AI) with one of his posts when the upvotes fell off and until the second one reached the top, with the intention of promoting his own blog. Let me simply prove my point to you on how predictable this spam is. He will do a blog post this month about this paper [0] with an expert analysis by either someone else (or even an LLM) with the primary intention of the blog being…

Paper and discussion history. A mention is not an endorsement.

mHC: Manifold-Constrained Hyper-Connections — HN submission

Thu, 01 Jan 2026 07:58:55 GMT

MHC: Manifold-Constrained Hyper-Connections

32 points; 3 discussion replies on HN. Paper and source history. A submission is not an endorsement.