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-2-2B
    4. Residual Stream - 16k
    5. 18-GEMMASCOPE-RES-16K
    6. 790
    Prev
    Next
    INDEX
    Explanations

    phrases related to deprivation and removal of rights or resources

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

    Something is being removed or taken away

    np_acts-logits-general · gemini-2.0-flash

    stripped of, cleared of, depleted of

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-2B @ 18-gemmascope-res-16k
    Configuration
    google/gemma-scope-2b-pt-res/layer_18/width_16k/average_l0_74
    Prompts (Dashboard)
    36,864 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    16,384
    Data Type
    float32
    Hook Name
    blocks.18.hook_resid_post
    Hook Layer
    18
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    igshid
    -0.65
    tvguidetime
    -0.63
     disambiguazione
    -0.57
    herzog
    -0.56
    initializeApp
    -0.56
    codable
    -0.55
    rungsseite
    -0.53
    becue
    -0.52
    istible
    -0.51
    Vanjske
    -0.51
    POSITIVE LOGITS
     cleans
    0.75
     rid
    0.71
     clean
    0.71
     stripped
    0.70
    clean
    0.65
     devoid
    0.65
     stripping
    0.65
    Clean
    0.63
    cleaned
    0.62
     Clean
    0.62
    Activations Density 0.180%

    No Known Activations