© 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 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. 17808
    Prev
    Next
    INDEX
    Explanations

    text related to conflicts or violent events

    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
    +0.17
    0.8%
    218
    +0.13
    0.6%
    530
    +0.12
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    326
    +0.17
    0.04
    530
    +0.13
    0.04
    659
    +0.12
    0.04
    Negative Logits
    etsk
    -0.68
    gio
    -0.62
    hran
    -0.61
     Reader
    -0.59
    aday
    -0.59
    mass
    -0.58
     Cipher
    -0.57
    hens
    -0.57
     Disciple
    -0.56
     pretext
    -0.55
    POSITIVE LOGITS
    OUN
    1.32
    ILL
    1.28
    OWN
    1.25
    EST
    1.24
    IZ
    1.23
    ALE
    1.22
     EDITION
    1.22
    ACTION
    1.21
    ARE
    1.21
    AND
    1.20
    Activations Density 0.130%

    No Known Activations