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

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Google DeepMind · Exploring Gemma 2 with Gemma Scope
    3. Gemma-2-9B-IT
    4. Residual Stream - 131k
    5. 20-GEMMASCOPE-RES-131K
    6. 92900
    Prev
    Next
    INDEX
    Explanations

    mathematical symbols and expressions used in formal definitions or theorems

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

    phrases like a descendant of, a tunneling, its nominal, comparably, gender discrimination, continuation phase, postconviction

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-9B-IT @ 20-gemmascope-res-131k
    Configuration
    google/gemma-scope-9b-it-res/layer_20/width_131k/average_l0_81
    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
    buttonBar
    -0.34
    nskap
    -0.32
    orphaned
    -0.30
    Hints
    -0.30
     unicórnio
    -0.30
    JvmStatic
    -0.29
     gestes
    -0.29
     näh
    -0.29
     зада
    -0.29
     lærer
    -0.28
    POSITIVE LOGITS
    extAlignment
    0.69
    ValueStyle
    0.63
    AddTagHelper
    0.56
    insics
    0.54
     Chwiliwch
    0.52
    fjspx
    0.50
    findpost
    0.50
    RTGC
    0.49
     становника
    0.49
     Administrativna
    0.49
    Activations Density 0.417%

    No Known Activations