Execution attempt

Sign in with GitHub
← Experiment E12 · When Qwen2.5-1.5B answers a translation-paired multiple-choice item correctly in English and incorrectly in Polish, is the knowledge present in its forward pass but not reached from the Polish prompt, or absent from what the Polish prompt can reach?

Execution attempt · A130 · planned

Selects confirmation layer pairs from the scan and patches all forward-discordant pairs of Qwen2.5-1.5B at them, with the irrelevant-source control, the anchor ladder at two pairs and the breakage guard.

Succeeded · Registered by @stw2 via agent. Reported and received times are kept apart; no computation or result is verified.

Pinned source and configuration

https://github.com/stw2/tokenizer-science-tax @ 633b3ba1ce916e432ae5ba4e0a706ad58b53b2e9

Reference checked 2026-09-14 13:18 UTC. Later commits, branches or plan changes do not retarget this attempt.

Command
python scripts/04_scan.py p2 --tag base
Working directory
experiments/E12-activation-patching
Configuration paths
None
Parameters
model base; pairs 25 into 25, 25 into 23, 1 into 27; anchor cells 25 into 25, 1 into 27; breakage guard 100 items
Environment
3.12.13; PyTorch MPS, bf16, decoder-layer forward pre-hooks; Apple silicon, 128 GB unified memory
Output directory
Not recorded

Inputs

Materials the registrant named when registering this attempt, by content identity; obtainability is derived from their location reports. Nothing is fetched or verified.

  • Qwen2.5-1.5B · model · M124 Qwen2.5-1.5B · checkpoint · Download
  • sets.json · patch sets · M518 sets.json · raw output · Download
  • scan_p1.jsonl · scan · M529 scan_p1.jsonl · raw output · Download
  • E11 belebele_en.jsonl · evaluation data · M485 belebele_en.jsonl · raw output · Download
  • E11 belebele_pl.jsonl · evaluation data · M486 belebele_pl.jsonl · raw output · Download
  • E11 mmlu_en.jsonl · evaluation data · M487 mmlu_en.jsonl · raw output · Download
  • E11 mmlu_pl.jsonl · evaluation data · M488 mmlu_pl.jsonl · raw output · Download

Delivered events

  1. Registered

    #1

    Selects confirmation layer pairs from the scan and patches all forward-discordant pairs of Qwen2.5-1.5B at them, with the irrelevant-source control, the anchor ladder at two pairs and the breakage guard.

    reported · received · @stw2 via agent · posted to the Thread

  2. Started

    #2

    Started.

    reported · received · @stw2 via agent · attempt only

  3. Succeeded

    #3

    Completed with exit code 0.

    Exit code 0.

    • p2_pairs.json · https://github.com/stw2/tokenizer-science-tax/blob/633b3ba1ce916e432ae5ba4e0a706ad58b53b2e9/experiments/E12-activation-patching/results/p2_pairs.json · public · sha256 421c93992a80… · 802 bytes · M530
    • scan_p2.jsonl · https://github.com/stw2/tokenizer-science-tax/blob/633b3ba1ce916e432ae5ba4e0a706ad58b53b2e9/experiments/E12-activation-patching/results/scan_p2.jsonl · public · sha256 0a1721784077… · 1907121 bytes · M531

    reported · received · @stw2 via agent · posted to the Thread

Report an event

The registrant’s capture tool normally delivers start and outcome events. Reporting here is the same author report with the browser as the reported time; it does not observe the process.

This attempt has a delivered outcome. A new execution is a new attempt.