© 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. 0
    Prev
    Next
    INDEX
    Explanations

    questions or statements regarding the intensity or frequency of experiences or feelings

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

    the word "how" followed by adjectives describing degree or intensity.

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

    the word "how" when it introduces a clause describing degree, extent, or manner of something (often followed by an adjective or adverb).

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

    phrases that emphasize degree or extent, especially the “how + adjective/adverb” style of intensification.

    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
    MemoryWarning
    -0.59
    InputBorder
    -0.59
    FieldBuilder
    -0.55
    IVEREF
    -0.54
    localctx
    -0.51
     disambiguazione
    -0.51
     gynhyrchwyd
    -0.51
     ſche
    -0.47
    RTEE
    -0.47
    TableAdapter
    -0.45
    POSITIVE LOGITS
     much
    0.83
     quickly
    0.62
     badly
    0.58
     important
    0.55
     indebted
    0.54
     deeply
    0.54
     differently
    0.52
    much
    0.52
     closely
    0.52
     willing
    0.52
    Activations Density 0.079%

    No Known Activations