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

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    1. Home
    3. Google DeepMind · Exploring Gemma 2 with Gemma Scope
    5. Gemma-2-9B
    7. Residual Stream - 16k
    9. 31-GEMMASCOPE-RES-16K
    11. 3417
    Prev
    Next
    INDEX
    Explanations

    contrastive conjunctions and terms indicative of hesitation or exception

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

    the word "but" used as a contrast or concession marker in sentences.

    oai_token-act-pair · claude-4-5-haikuTriggered by @emiglarou
    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
    fromnode
    -0.68
    adaptiveStyles
    -0.64
    Cyfarwyddwr
    -0.62
     wikipagina
    -0.61
    RenderAtEndOf
    -0.60
    jspb
    -0.60
    الدراسه
    -0.60
    WireFormatLite
    -0.57
    elemField
    -0.56
    لينكات
    -0.56
    POSITIVE LOGITS
     I
    0.71
     honestly
    0.62
     it
    0.58
     there
    0.57
     remember
    0.57
     you
    0.57
    I
    0.56
     we
    0.53
     realistically
    0.53
     they
    0.52
    Activations Density 0.218%

    No Known Activations