© 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
    2. Google DeepMind · Exploring Gemma 2 with Gemma Scope
    3. Gemma-2-9B
    4. Residual Stream - 16k
    5. 31-GEMMASCOPE-RES-16K
    6. 5844
    Prev
    Next
    INDEX
    Explanations

    language-related constructs, particularly in programming and code formatting

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

    French language text, particularly French pronouns and verb conjugations.

    oai_token-act-pair · claude-4-5-haikuTriggered by @emiglarou

    text written in Romance languages, particularly French, Spanish, and Portuguese.

    oai_token-act-pair · claude-4-5-sonnetTriggered by @jyhe0408

    non-English Romance-language text fragments, especially verb and negation morphemes with accented characters.

    oai_token-act-pair · gpt-5Triggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-9B @ 31-gemmascope-res-16k
    Configuration
    google/gemma-scope-9b-pt-res/layer_31/width_16k/average_l0_114
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    16,384
    Data Type
    float32
    Hook Name
    blocks.31.hook_resid_post
    Hook Layer
    31
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    ItemBackground
    -0.55
     serem
    -0.53
     betweenstory
    -0.51
    TargetException
    -0.51
    URLException
    -0.48
    ChildScrollView
    -0.48
     Autorizaciones
    -0.47
     незавершена
    -0.45
     undone
    -0.44
     ***!
    -0.43
    POSITIVE LOGITS
     is
    0.71
     has
    0.66
     does
    0.61
    PYX
    0.54
     tends
    0.54
     can
    0.52
     hasn
    0.50
     isn
    0.50
     seems
    0.49
    volves
    0.46
    Activations Density 0.214%

    No Known Activations