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

    Neuronpedia

    Natural Language
    Autoencoders
    NEW
    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    2. Joseph Bloom · Initial Gemma-2B SAEs
    3. Gemma-2B
    4. Residual Stream
    5. 0-RES-JB
    6. 976
    Prev
    Next
    INDEX
    Explanations

    names or references to a specific person named "Derek"

    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
    1806
    +0.09
    0.3%
    645
    +0.08
    0.3%
    1323
    +0.08
    0.3%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    129
    +0.09
    0.04
    1575
    +0.08
    0.03
    690
    +0.08
    0.03
    Negative Logits
    <?
    -0.77
     bus
    -0.69
     catch
    -0.68
    bus
    -0.67
     transit
    -0.66
     go
    -0.65
     set
    -0.65
     do
    -0.64
    public
    -0.64
     policy
    -0.64
    POSITIVE LOGITS
     accla
    2.02
     increa
    2.00
     impra
    1.94
     inev
    1.92
     disagre
    1.88
     unwarran
    1.88
     affor
    1.80
     depic
    1.79
     viciss
    1.78
     wherea
    1.77
    Activations Density 0.166%

    No Known Activations