© 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 49152
    9. 8-RES_FS49152-JB
    11. 43120
    Prev
    Next
    INDEX
    Explanations

    mentions of individuals' titles and academic or professional designations

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs49152-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_49152
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    49,152
    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
    431
    +0.12
    0.6%
    180
    +0.11
    0.5%
    376
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    431
    +0.12
    0.03
    180
    +0.11
    0.02
    376
    +0.11
    0.02
    Negative Logits
     comprom
    -0.78
    theless
    -0.69
     festive
    -0.64
     divid
    -0.62
     shards
    -0.62
    =-=-=-=-=-=-=-=-
    -0.61
    keeping
    -0.61
     DIY
    -0.59
    schild
    -0.59
     simmer
    -0.59
    POSITIVE LOGITS
    .,
    1.50
    .;
    1.13
    arcer
    1.12
    .?
    1.11
    ./
    1.10
    .,"
    1.05
    .:
    1.05
    orporated
    1.03
    .—
    0.97
     Jr
    0.93
    Activations Density 0.010%

    No Known Activations