© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Joseph Bloom · Initial Gemma-2B SAEs
    5. Gemma-2B
    7. Residual Stream
    9. 0-RES-JB
    11. 2181
    Prev
    Next
    INDEX
    Explanations

    finger-related terms or situations involving fingers being injured

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    jbloom/Gemma-2b-Residual-Stream-SAEs/gemma_2b_blocks.0.hook_resid_post_16384_anthropic
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    HuggingFaceFW/fineweb
    Features
    16,384
    Data Type
    torch.float32
    Hook Point
    blocks.0.hook_resid_post
    Architecture
    standard
    Context Size
    1,024
    Dataset
    HuggingFaceFW/fineweb
    Hook Point Layer
    0
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Neuron Alignment
    Index
    Value
    % of L₁
    1926
    +0.07
    0.2%
    1310
    +0.07
    0.2%
    1602
    +0.07
    0.2%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    1910
    +0.07
    0.03
    330
    +0.07
    0.03
    201
    +0.07
    0.03
    Negative Logits
    <bos>
    -1.44
    public
    -0.82
    -0.79
    protected
    -0.79
    BUYER
    -0.78
    static
    -0.77
    }{||
    -0.77
    void
    -0.77
    <tr>
    -0.76
    CreateMap
    -0.76
    POSITIVE LOGITS
     increa
    2.87
     maneu
    2.76
     affor
    2.73
     fta
    2.68
     impra
    2.60
     stockholm
    2.58
     inev
    2.54
     disagre
    2.51
     depic
    2.50
     aen
    2.50
    Activations Density 0.136%
    No Known Activations