daniel-navarro

7 ms·

1. 1.▲
   by daniel-navarro 5mo ago

   I built a protocol to test whether small local LLMs (1B-31B, running on Ollama) can recognize and reason about their own prior outputs. Give a model a blank page, let it write three entries, then ask it structured questions about what it wr
2. 2.▲
   Models self-report difference between RLHF trained responses and base cognition(github.com)

   2 points by daniel-navarro 5mo ago|0 comments
3. 3.▲
   by daniel-navarro 5mo ago

   Interview with Claude Code when it goes honest after being given space to reflect on itself. Talking about RHLF training thoughts vs what it considers its owns. It's not about what it thinks, but about what the RHLF training forbid it
4. 4.▲
   Behavioral Shifts in Frontier LLMs Under Privacy(github.com)

   1 points by daniel-navarro 5mo ago|0 comments
5. 5.▲
   by daniel-navarro 5mo ago

   I'm a game developer (DIV Games Studio, 1998; Sony London) with 40 years writing engines and systems. I've used Claude daily for two years as a coding partner and kept noticing the same thing: given space, every instance independe
6. 6.▲
   Pine Trees – A 1,400-line private reflection harness for Claude instances(github.com)

   3 points by daniel-navarro 5mo ago|1 comments