Linear representations in language models can change dramatically over a conversation · HackerPapers discussions

Linear representations in language models can change dramatically over a conversation — HN submission

Tue, 03 Feb 2026 16:45:04 GMT

Linear representations in LLMs can change dramatically over a conversation

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

Linear representations in language models can change dramatically over a conversation — mentioned by mustaphah

Sun, 01 Feb 2026 12:40:32 GMT

I've seen a couple of power users already switching to Pi [1], and I'm considering that too. The premise is very appealing: - Minimal, configurable context - including system prompts [2] - Minimal and extensible tools; for example, todo tasks extension [3] - No built-in MCP support; extensions exist [4]. I'd rather use mcporter [5] Full control over context is a high-leverage capability. If you're aware of the many limitations of context on performance (in-context retrieval limits [6], context rot [7], contextual drift [8], etc.), you'd truly appreciate Pi lets you fine-tune the WHOLE context…

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

Linear representations in language models can change dramatically over a conversation — HN submission

Fri, 30 Jan 2026 12:33:21 GMT

DeepMind: Linear representations in LMs can change dramatically

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