© 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. 6050
    Prev
    Next
    INDEX
    Model
    gemma-2-9b-it
    Layer #
    20
    Steering Hook
    blocks.20.hook_resid_pre
    Steering Strength
    50.25
    Uploader
    bot-neuronpedia
    Created At
    2/15/2025 1:06:43 AM
    Raw Vector
    Actions
    Explanations

    the presence of full stops and punctuation that indicates a conclusion or end of thought

    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
     
    -0.32
     fertiliser
    -0.32
    erals
    -0.31
    ījum
    -0.30
     Heimat
    -0.30
    亮相
    -0.30
     infrastruktur
    -0.30
    Request
    -0.29
    erobic
    -0.29
     Aurelius
    -0.29
    POSITIVE LOGITS
    setVerticalGroup
    0.65
     betweenstory
    0.65
     MessageBoxIcon
    0.63
     kasarigan
    0.59
     queſta
    0.58
     Taktlose
    0.57
     resourceCulture
    0.57
     שוליים
    0.56
    fromnode
    0.56
    ########.
    0.55
    Activations Density 0.014%

    No Known Activations