© 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. 2383
    Prev
    Next
    INDEX
    Explanations

    mentions of specific individuals, records, and officials, as well as verbs related to actions taken by individuals

    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
    480
    +0.44
    2.3%
    87
    +0.16
    0.8%
    266
    +0.13
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    480
    +0.44
    0.03
    266
    +0.16
    0.09
    288
    +0.13
    0.10
    Negative Logits
    ITNESS
    -0.69
    ifted
    -0.62
     confir
    -0.60
     notor
    -0.60
     incorpor
    -0.56
     Moroc
    -0.56
     comr
    -0.54
     opio
    -0.53
     reluct
    -0.53
     pestic
    -0.53
    POSITIVE LOGITS
     onto
    0.74
     into
    0.66
     goodbye
    0.57
    til
    0.55
     throats
    0.55
     itch
    0.55
    orts
    0.55
     Into
    0.54
     hostage
    0.53
     unnecessarily
    0.52
    Activations Density 0.869%

    No Known Activations