Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Google DeepMind · Exploring Gemma 2 with Gemma Scope
    3. Gemma-2-9B-IT
    4. Residual Stream - 16k
    5. 31-GEMMASCOPE-RES-16K
    6. 15422
    Prev
    Next
    INDEX
    Explanations

    expressions of frustration and disbelief in personal situations

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

    Following conversational cues such as "Well," "oh," and "<start_of_turn>user"

    np_acts-logits-general · gemini-2.0-flash

    dialogue starting phrases

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-9B-IT @ 31-gemmascope-res-16k
    Configuration
    google/gemma-scope-9b-it-res/layer_31/width_16k/average_l0_76
    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
     pihaknya
    -0.36
     undangan
    -0.29
     ...
    -0.29
     alcuna
    -0.29
     délib
    -0.29
     itself
    -0.28
     dedans
    -0.28
     enige
    -0.27
     oczywiście
    -0.27
     personlig
    -0.27
    POSITIVE LOGITS
    出版年
    0.75
    BIÉN
    0.63
    Portale
    0.63
     betweenstory
    0.63
     Autorisations
    0.60
    !”.
    0.60
     mommy
    0.60
    〇〇
    0.59
     XYZ
    0.59
     iprot
    0.59
    Activations Density 0.205%

    No Known Activations