© 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. Google DeepMind · Exploring Gemma 2 with Gemma Scope
    5. Gemma-2-9B
    7. Residual Stream - 16k
    9. 20-GEMMASCOPE-RES-16K
    11. 3249
    Prev
    Next
    INDEX
    Explanations

    phrases that express personal opinions and claims

    oai_token-act-pair · gpt-4o-miniTriggered by @bot

    phrases with the word "say" and related reporting or declarative language structures.

    oai_token-act-pair · claude-3-5-haiku-20241022Triggered by @neilrathi

    constructions where a speaker frames or hedges an assertion using a reporting verb followed by a “that”-clause.

    oai_token-act-pair · gpt-5Triggered by @jyhe0408

    the phrase "say that" when introducing a statement, claim, or assertion.

    oai_token-act-pair · claude-4-5-sonnetTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-9B @ 20-gemmascope-res-16k
    Configuration
    google/gemma-scope-9b-pt-res/layer_20/width_16k/average_l0_68
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    16,384
    Data Type
    float32
    Hook Name
    blocks.20.hook_resid_post
    Hook Layer
    20
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     InputDecoration
    -0.56
    tagHelperRunner
    -0.52
     ſta
    -0.45
     purpoſe
    -0.44
     supuestamente
    -0.43
    principalColumn
    -0.42
    heav
    -0.42
     uſe
    -0.40
     pleaſure
    -0.40
    LEGGI
    -0.40
    POSITIVE LOGITS
    endphp
    0.56
    可以说
    0.54
     arguably
    0.49
    ChildScrollView
    0.46
    可以说是
    0.46
     glyphicon
    0.44
    Certainly
    0.42
    と言える
    0.42
     artikkeli
    0.41
     fairly
    0.41
    Activations Density 0.193%

    No Known Activations