© 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. 16963
    Prev
    Next
    INDEX
    Explanations

    organizations, locations, and events mentioned in an online fundraising or support 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
    481
    +0.17
    0.8%
    447
    +0.16
    0.7%
    314
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    288
    +0.17
    0.06
    232
    +0.16
    0.04
    537
    +0.10
    0.04
    Negative Logits
    Downloadha
    -0.81
     gif
    -0.71
    amsung
    -0.67
     silence
    -0.65
     predec
    -0.65
     conflic
    -0.65
     describ
    -0.63
     mic
    -0.62
     nausea
    -0.61
     toget
    -0.60
    POSITIVE LOGITS
    istries
    0.84
     Railway
    0.83
     Association
    0.81
    isine
    0.81
    Works
    0.81
    atorium
    0.79
    Fest
    0.79
    roots
    0.79
    ways
    0.78
     Consortium
    0.78
    Activations Density 0.306%

    No Known Activations