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. 7-GEMMASCOPE-RES-16K
    6. 2151
    Prev
    Next
    INDEX
    Explanations

    references to graphic or gruesome content

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

    keywords and phrases associated with horror and bodily grotesqueness

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-2B @ 7-gemmascope-res-16k
    Configuration
    google/gemma-scope-2b-pt-res/layer_7/width_16k/average_l0_69
    Prompts (Dashboard)
    36,864 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    16,384
    Data Type
    float32
    Hook Name
    blocks.7.hook_resid_post
    Hook Layer
    7
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     المعيارى
    -0.79
    AddTagHelper
    -0.70
    createCanvas
    -0.67
     AssemblyCompany
    -0.67
    tableFuture
    -0.66
    jsonwebtoken
    -0.65
    InputTagHelper
    -0.65
    stateProvider
    -0.65
     financieras
    -0.63
    IntoConstraints
    -0.61
    POSITIVE LOGITS
     disgusting
    1.19
     horrible
    1.16
     horror
    1.14
     horri
    1.07
     horrific
    1.06
     horrors
    1.05
     horrifying
    1.02
    horror
    1.01
     hideous
    1.00
     nasty
    0.98
    Activations Density 0.335%

    No Known Activations