© 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 3072
    9. 8-RES_FS3072-JB
    11. 731
    Prev
    Next
    INDEX
    Explanations

    words related to official titles or positions

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs3072-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_3072
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    3,072
    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
    128
    +0.11
    0.5%
    283
    +0.11
    0.5%
    23
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    232
    +0.11
    0.10
    288
    +0.11
    0.10
    128
    +0.11
    0.07
    Negative Logits
    Ø©
    -0.75
    ãģĦ
    -0.71
    OAD
    -0.68
    士
    -0.67
    ITED
    -0.66
    STON
    -0.65
    ãģį
    -0.60
    issance
    -0.60
    lly
    -0.59
    senal
    -0.59
    POSITIVE LOGITS
    ropy
    1.31
    rans
    1.29
    ypes
    1.24
    ral
    1.23
    rell
    1.22
    raction
    1.21
    race
    1.21
    ainment
    1.20
    rip
    1.20
    ract
    1.20
    Activations Density 1.095%

    No Known Activations