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

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsBlogSlackPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Google DeepMind · Exploring Gemma 2 with Gemma Scope
    3. Gemma-2-9B
    4. Residual Stream - 131k
    5. 20-GEMMASCOPE-RES-131K
    6. 57330
    Prev
    Next
    INDEX
    Explanations

    instructions or actions related to configuring and integrating software systems

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

    Technical action phrases that describe setup, configuration or implementation processes, typically appearing in documentation or tutorial content, often in the format of "how to [verb] X with/using Y".

    eleuther_acts_top20 · claude-3-7-sonnet-20250219Triggered by @lu-christina
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-9b-pt-res/layer_20/width_131k/average_l0_114
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    131,072
    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
     avoient
    -0.65
     enfans
    -0.64
    findpost
    -0.61
     macetas
    -0.54
     varandra
    -0.54
     auroit
    -0.49
     semmi
    -0.47
    anskje
    -0.47
     mitään
    -0.45
     PLWABN
    -0.44
    POSITIVE LOGITS
     Steps
    0.46
     Sted
    0.45
     yourself
    0.44
     sted
    0.43
    "}")
    0.43
     Fold
    0.42
     DIY
    0.41
     Riv
    0.40
     Vog
    0.40
     *
    
    
    0.40
    Activations Density 0.089%

    No Known Activations