Skip to content

Publish PiPER contact history v0.3 and all 84 simulation attempts - #28

Merged
Tarzelf merged 1 commit into
mainfrom
feat/piper-contact-memory-v03
Oct 12, 2026
Merged

Tarzelf merged 1 commit into
mainfrom
feat/piper-contact-memory-v03

Conversation

@Tarzelf

@Tarzelf Tarzelf commented Oct 12, 2026

Copy link
Copy Markdown
Contributor

Publish the frozen PiPER contact-history v0.3 comparison: history completes 6/12 placements versus 4/12 for both original and probe-only motion. All 84 native attempts are valid and all 48 negative controls remain unsuccessful. The two added completions are 60 g nominal-friction blocks; easier cases incur extra task time, and all six history acquisition retries still fail.

The paper and research homepage expose the complete original 90-file receipt, the frozen source, separate 13-attempt development archive, execution timing boundary and immutable Hugging Face evidence. Actual movement after requested slowdown/hold remains visible, including two intervals that initially accelerate, five resumes across four completed tasks, and one resume failing the independent low absolute-motion check. No physical sensing, learned policy or hardware qualification is claimed.

Validation covers all stored and decoded hashes, the full paired denominator, source/native model identity, exact matched open-jaw command tapes, causal recorded history and task outcomes. Runtime timing consists of per-attempt summaries, maxima and miss counts; raw per-decision wall-time samples are not retained.

Training code and evidence PR · Hugging Face original bundle

Validation: npm run check passed; all 46 research tests passed. The site includes 17 papers and 605 public artifacts.

@Tarzelf
Tarzelf merged commit cb52c45 into main Oct 12, 2026
1 check passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant