© 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. 24557
    Prev
    Next
    INDEX
    Explanations

    words that express moderate degrees or levels of something

    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
    195
    +0.11
    0.5%
    151
    +0.10
    0.5%
    551
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    331
    +0.11
    0.01
    151
    +0.10
    0.01
    221
    +0.10
    0.01
    Negative Logits
    abad
    -0.87
    ULTS
    -0.86
    yrs
    -0.80
    ovies
    -0.78
    mson
    -0.77
    ults
    -0.76
    aring
    -0.74
    yers
    -0.71
    olan
    -0.71
     Delivery
    -0.70
    POSITIVE LOGITS
     amusing
    0.93
     confusing
    0.87
     surprising
    0.86
     ironic
    0.85
     obscure
    0.83
     understandable
    0.83
     neb
    0.81
     ambiguous
    0.81
     resembling
    0.81
     predictable
    0.80
    Activations Density 0.013%

    No Known Activations