© 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. 22362
    Prev
    Next
    INDEX
    Explanations

    terms associated with influence and the role of various entities in context

    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
    236
    +0.11
    0.5%
    497
    +0.11
    0.5%
    284
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    236
    +0.11
    0.04
    285
    +0.11
    0.03
    284
    +0.10
    0.04
    Negative Logits
    ched
    -0.61
    achelor
    -0.61
    é¾
    -0.61
    Catalog
    -0.60
    Seriously
    -0.60
    swick
    -0.58
    ignore
    -0.58
    APTER
    -0.57
    chers
    -0.57
    intend
    -0.57
    POSITIVE LOGITS
    iveness
    0.78
     inherent
    0.72
     exerted
    0.70
    effects
    0.70
     of
    0.69
    arth
    0.68
     Painter
    0.67
    lessness
    0.67
     factor
    0.65
     dynamic
    0.65
    Activations Density 0.118%

    No Known Activations