© 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. 24386
    Prev
    Next
    INDEX
    Explanations

    verbs associated with managing, solving problems, and engaging in tasks

    oai_token-act-pair · gpt-4o-miniTriggered 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
    87
    +0.21
    1.0%
    665
    +0.11
    0.6%
    756
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    665
    +0.21
    0.09
    570
    +0.11
    0.03
    70
    +0.10
    0.07
    Negative Logits
    itcher
    -0.65
    DragonMagazine
    -0.62
    Serial
    -0.56
    iasco
    -0.54
     Deadline
    -0.54
    esides
    -0.53
     fortunately
    -0.53
    brainer
    -0.52
    Enlarge
    -0.51
    mpeg
    -0.50
    POSITIVE LOGITS
     oneself
    1.12
     them
    0.93
    enance
    0.84
     themselves
    0.80
     anything
    0.75
     whatever
    0.75
    urate
    0.74
     somebody
    0.74
     someone
    0.73
     him
    0.72
    Activations Density 0.485%

    No Known Activations