Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsBlogSlackPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Joseph Bloom · Feature Splitting for GPT2-Small
    3. GPT2-Small
    4. Residual Feature Splitting 768
    5. 8-RES_FS768-JB
    6. 320
    Prev
    Next
    INDEX
    Explanations

    words related to motion or directional movement

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs768-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_768
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    768
    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
    354
    +0.12
    0.6%
    377
    +0.12
    0.6%
    437
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    354
    +0.12
    0.11
    377
    +0.12
    0.12
    200
    +0.10
    0.10
    Negative Logits
    ullah
    -0.79
    ament
    -0.76
    creen
    -0.68
     Horus
    -0.64
    eers
    -0.64
    ificent
    -0.62
    icio
    -0.61
    oret
    -0.58
    idable
    -0.57
    lake
    -0.56
    POSITIVE LOGITS
     overboard
    1.04
    verning
    1.04
    lems
    0.99
    vt
    0.99
    ggle
    0.96
     viral
    0.88
     unnoticed
    0.84
    Ń·
    0.82
     bankrupt
    0.81
     extinct
    0.79
    Activations Density 2.193%

    No Known Activations