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

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsBlogSlackPrivacy & 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. 215
    Prev
    Next
    INDEX
    Explanations

    sensitive topics related to sexual assault, abuse, and misconduct

    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
    384
    +0.10
    0.4%
    484
    +0.10
    0.4%
    235
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    288
    +0.10
    0.16
    235
    +0.10
    0.11
    325
    +0.10
    0.10
    Negative Logits
    ocument
    -0.85
     Flavoring
    -0.84
    BLIC
    -0.78
    arily
    -0.77
    IVERS
    -0.75
    REC
    -0.75
     Solitaire
    -0.74
     GOODMAN
    -0.68
    Dispatch
    -0.68
    overed
    -0.67
    POSITIVE LOGITS
    volent
    1.16
     ejac
    1.06
     genital
    1.00
     condom
    0.99
     sex
    0.99
     genitals
    0.98
    ager
    0.97
     condoms
    0.93
     underwear
    0.91
     rape
    0.89
    Activations Density 4.128%

    No Known Activations