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. 2658
    Prev
    Next
    INDEX
    Explanations

    sequences of alphanumeric characters or variations of the same string

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

    abcdef sequences and variations

    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
    verifyException
    -0.42
    väg
    -0.38
    ⓘ
    -0.37
     culte
    -0.37
    ]")]
    -0.36
     tip
    -0.36
     amid
    -0.35
    £
    -0.35
    $")
    -0.35
     hearers
    -0.35
    POSITIVE LOGITS
    abcdefghijklmnop
    0.98
    GHIJKLM
    0.98
    qrstuvwxyz
    0.91
    UVWXYZ
    0.85
    NOPQRST
    0.84
     فريبيس
    0.79
    ABCDEF
    0.73
    WriteTagHelper
    0.73
    titleMargin
    0.63
    abcdef
    0.63
    Activations Density 0.006%

    No Known Activations