© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-9B-IT
    5. 20-AXBENCH-REFT-R1-RES-16K
    7. 4566
    Prev
    Next
    INDEX
    Model
    gemma-2-9b-it
    Layer #
    20
    Steering Hook
    blocks.20.hook_resid_pre
    Steering Strength
    49.5
    Uploader
    bot-neuronpedia
    Created At
    2/15/2025 1:06:43 AM
    Raw Vector
    Actions
    Explanations

    punctuation marks, particularly periods, quotes, and numbers

    oai_token-act-pair · gpt-4o-mini
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    pyvene/gemma-reft-r1-9b-it-res/l20
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    No Configuration Found
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     responsible
    -0.39
    Personendaten
    -0.38
     responsable
    -0.32
     tua
    -0.32
     traditionally
    -0.31
     hydraulic
    -0.30
     Kommune
    -0.30
     nawr
    -0.30
    ).__
    -0.30
    Request
    -0.29
    POSITIVE LOGITS
    setVerticalGroup
    0.62
     betweenstory
    0.60
     defaultstate
    0.57
     queſta
    0.55
    LookAnd
    0.54
    expandindo
    0.54
    HandlerContext
    0.54
    +#+#
    0.53
     Taktlose
    0.52
     purpoſe
    0.52
    Activations Density 0.697%

    No Known Activations