© 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. 13360
    Prev
    Next
    INDEX
    Explanations

    phrases related to controversy and discussion around societal issues

    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
    608
    +0.16
    0.7%
    373
    +0.11
    0.5%
    549
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    608
    +0.16
    0.07
    549
    +0.11
    0.04
    669
    +0.11
    0.04
    Negative Logits
    awei
    -0.72
    sky
    -0.67
    DragonMagazine
    -0.63
    tun
    -0.63
     redirected
    -0.62
    DH
    -0.62
     indexed
    -0.59
     phased
    -0.59
     coupled
    -0.59
     elig
    -0.59
    POSITIVE LOGITS
    ggles
    0.92
    wered
    0.91
     appease
    0.90
     accompany
    0.90
     ensure
    0.88
     avoid
    0.80
     accommodate
    0.78
     achieve
    0.78
     commemorate
    0.77
     coincide
    0.76
    Activations Density 0.183%

    No Known Activations