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. 16366
    Prev
    Next
    INDEX
    Explanations

    variations of quotation marks and other punctuation used for quotes or dialogue

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

    Text enclosed in quotation marks

    np_acts-logits-general · gemini-2.0-flash

    quoted terms

    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
      
    -0.27
       
    -0.27
     /
    -0.26
     with
    -0.25
        
    -0.25
    「
    -0.23
     \
    -0.23
     and
    -0.23
     is
    -0.22
    	
    -0.22
    POSITIVE LOGITS
    niſſe
    1.28
    ſſung
    1.27
    iſchen
    1.27
     $_-
    1.27
    <unused16>
    1.25
    <unused14>
    1.25
    <unused17>
    1.25
    <unused23>
    1.25
    [@BOS@]
    1.25
    <unused8>
    1.25
    Activations Density 0.285%

    No Known Activations