© 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. 22492
    Prev
    Next
    INDEX
    Explanations

    references to mobile technology and gaming elements

    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
    447
    +0.22
    1.1%
    281
    +0.11
    0.5%
    275
    +0.09
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    288
    +0.22
    0.09
    537
    +0.11
    0.07
    281
    +0.09
    0.04
    Negative Logits
    ãĤ¨ãĥ«
    -0.67
    sonian
    -0.60
    usalem
    -0.57
    æ©Ł
    -0.57
    hene
    -0.56
    EGIN
    -0.56
    cedented
    -0.54
    itizens
    -0.54
    é¾įåĸļ士
    -0.53
    iannopoulos
    -0.53
    POSITIVE LOGITS
     whereby
    0.77
     which
    0.70
     Lastly
    0.67
    !:
    0.65
     (-
    0.59
     (.
    0.58
     wherein
    0.57
     (>
    0.55
    ossibility
    0.54
     besides
    0.53
    Activations Density 0.974%

    No Known Activations