© 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. 10212
    Prev
    Next
    INDEX
    Explanations

    research-related terms, such as researchers, team, group, study, and analysis

    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
    610
    +0.13
    0.6%
    447
    +0.12
    0.6%
    481
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    610
    +0.13
    0.03
    346
    +0.12
    0.03
    403
    +0.11
    0.03
    Negative Logits
     Canaveral
    -0.73
    Bound
    -0.71
     Vid
    -0.70
    Lago
    -0.70
     Maiden
    -0.68
    ãĥĸ
    -0.68
    VW
    -0.66
     sunrise
    -0.66
     decency
    -0.65
    minster
    -0.65
    POSITIVE LOGITS
     hypothes
    1.05
     authors
    1.02
     concluded
    0.97
     hypothesized
    0.97
     collaborated
    0.96
     analyzed
    0.95
     conclude
    0.95
     recruited
    0.93
     speculate
    0.92
     studied
    0.92
    Activations Density 0.065%

    No Known Activations