© 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 · Feature Splitting for GPT2-Small
    5. GPT2-Small
    7. Residual Feature Splitting 24576
    9. 8-RES_FS24576-JB
    11. 24383
    Prev
    Next
    INDEX
    Explanations

    expressions of gratitude and appreciation towards influential individuals

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs24576-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_24576
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    24,576
    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
    32
    +0.12
    0.5%
    218
    +0.11
    0.5%
    735
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    105
    +0.12
    0.05
    467
    +0.11
    0.05
    262
    +0.10
    0.04
    Negative Logits
    enum
    -0.74
    ptive
    -0.73
    atten
    -0.68
     modifier
    -0.67
     defaults
    -0.67
    ivably
    -0.67
    CLAIM
    -0.66
    essim
    -0.65
     Worse
    -0.64
    Elsewhere
    -0.64
    POSITIVE LOGITS
     generosity
    1.28
     invaluable
    1.28
     kindness
    1.28
     gracious
    1.25
     generous
    1.17
     professionalism
    1.10
     generously
    1.09
     kindly
    1.07
     thoughtful
    1.04
     THANK
    1.03
    Activations Density 1.008%

    No Known Activations