© 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
    3. Joseph Bloom · Feature Splitting for GPT2-Small
    5. GPT2-Small
    7. Residual Feature Splitting 49152
    9. 8-RES_FS49152-JB
    11. 26811
    Prev
    Next
    INDEX
    Explanations

    occurrences of the word "diverge" or its variations, indicating a focus on concepts related to divergence or differences

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs49152-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_49152
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    49,152
    Data Type
    torch.float32
    Hook Point
    blocks.8.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    8
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Neuron Alignment
    Index
    Value
    % of L₁
    0
    +0.14
    0.6%
    232
    +0.11
    0.5%
    537
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    0
    +0.14
    0.01
    624
    +0.11
    0.01
    232
    +0.11
    0.01
    Negative Logits
    EEK
    -0.81
    è¦ļéĨĴ
    -0.74
     Introduced
    -0.70
    ORED
    -0.67
    Fax
    -0.66
    ORN
    -0.63
    ammy
    -0.63
     toast
    -0.61
    OW
    -0.60
    HOU
    -0.60
    POSITIVE LOGITS
    gent
    1.41
    gencies
    1.30
    tic
    1.19
    ging
    1.16
    gently
    1.12
    gence
    1.12
    gency
    1.08
    ministic
    1.08
    ting
    1.04
    ged
    0.99
    Activations Density 0.006%

    No Known Activations