© 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 24576
    9. 8-RES_FS24576-JB
    11. 2227
    Prev
    Next
    INDEX
    Explanations

    mentions of activities or situations outside of a certain context

    oai_token-act-pair · gpt-3.5-turboTriggered 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
    373
    +0.10
    0.5%
    662
    +0.10
    0.5%
    697
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    662
    +0.10
    0.02
    697
    +0.10
    0.02
    505
    +0.10
    0.02
    Negative Logits
    enegger
    -0.86
    _-
    -0.79
    anwhile
    -0.79
    lda
    -0.77
    ruary
    -0.74
    iji
    -0.73
    dayName
    -0.72
    amaz
    -0.68
    arrell
    -0.67
     TAMADRA
    -0.66
    POSITIVE LOGITS
     world
    0.87
     observer
    0.86
    world
    0.80
     workings
    0.77
    shock
    0.71
     linebacker
    0.71
     linebackers
    0.71
     attractiveness
    0.69
     hemisphere
    0.69
     diameter
    0.69
    Activations Density 0.032%

    No Known Activations