© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    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. 20346
    Prev
    Next
    INDEX
    Explanations

    locations and specific places

    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
    564
    +0.12
    0.6%
    168
    +0.11
    0.5%
    693
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    693
    +0.12
    0.05
    708
    +0.11
    0.04
    564
    +0.10
    0.04
    Negative Logits
    adapt
    -0.79
    laus
    -0.77
    onal
    -0.74
    layer
    -0.73
    rating
    -0.71
    GM
    -0.71
    schild
    -0.70
    sam
    -0.70
    product
    -0.69
    itic
    -0.68
    POSITIVE LOGITS
     Correspond
    1.01
     courthouse
    0.99
     podium
    0.93
     flanked
    0.93
     desks
    0.90
    rooms
    0.84
     Building
    0.83
     hallway
    0.81
     entrances
    0.81
     ceremony
    0.81
    Activations Density 0.129%

    No Known Activations