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 - 16k
    5. 20-GEMMASCOPE-RES-16K
    6. 827
    Prev
    Next
    INDEX
    Explanations

    references to locks, keys, and related security mechanisms

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

    Lock, key, password-related words

    np_acts-logits-general · gemini-2.0-flash

    keys and locks

    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-16k
    Configuration
    google/gemma-scope-9b-it-res/layer_20/width_16k/average_l0_91
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    16,384
    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
     Cover
    -0.39
    Invariant
    -0.38
     Indirect
    -0.35
     محفوظة
    -0.35
     Confront
    -0.35
    Cover
    -0.34
     cover
    -0.34
    httphttps
    -0.33
    ELER
    -0.31
     gewinnen
    -0.31
    POSITIVE LOGITS
     locksmith
    0.65
     виправивши
    0.64
    ChildScrollView
    0.62
     betweenstory
    0.60
     OMITBAD
    0.59
    principalColumn
    0.58
     resourceCulture
    0.53
    PreferredItem
    0.53
    \{\\
    0.50
     المعيارى
    0.50
    Activations Density 0.145%

    No Known Activations