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

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Joseph Bloom · Initial Gemma-2B-IT SAEs
    3. Gemma-2B-IT
    4. Residual Stream
    5. 12-RES-JB
    6. 8346
    Prev
    Next
    INDEX
    Explanations

    names or variables that are abbreviated using the format initial<tab>capitalized

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    jbloom/Gemma-2b-IT-Residual-Stream-SAEs/gemma_2b_it_blocks.12.hook_resid_post_16384
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    HuggingFaceFW/fineweb
    Features
    16,384
    Data Type
    float32
    Hook Name
    blocks.12.hook_resid_post
    Hook Layer
    12
    Architecture
    standard
    Context Size
    1,024
    Dataset
    Skylion007/openwebtext
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Neuron Alignment
    Index
    Value
    % of L₁
    1177
    +0.12
    0.4%
    964
    +0.12
    0.4%
    453
    +0.12
    0.3%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    981
    +0.12
    0.06
    1001
    +0.12
    0.05
    199
    +0.12
    0.03
    Negative Logits
    <bos>
    -0.76
    .
    -0.60
    !
    -0.59
     and
    -0.59
     a
    -0.58
     or
    -0.58
     of
    -0.58
     (
    -0.58
     s
    -0.57
     to
    -0.57
    POSITIVE LOGITS
     alkoh
    1.75
     kram
    1.67
     makro
    1.61
     silikon
    1.58
     keramik
    1.56
     uhr
    1.55
     antik
    1.53
     maksi
    1.53
     kac
    1.53
     kompati
    1.52
    Activations Density 0.124%

    No Known Activations