© 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. 9980
    Prev
    Next
    INDEX
    Explanations

    phrases related to specific named entities (e.g., names of people or locations)

    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
    195
    +0.12
    0.6%
    314
    +0.12
    0.5%
    746
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    746
    +0.12
    0.01
    603
    +0.12
    0.01
    702
    +0.11
    0.01
    Negative Logits
     Hercules
    -0.79
     Prometheus
    -0.73
     Pixie
    -0.73
     Wester
    -0.66
     Gemini
    -0.65
     incidentally
    -0.64
     Bayer
    -0.64
     Gauntlet
    -0.64
     Fundamental
    -0.63
     Paradox
    -0.63
    POSITIVE LOGITS
    mur
    1.34
    geon
    1.09
    ricane
    1.03
    rats
    1.02
    phys
    1.02
    abit
    1.00
    asaki
    1.00
    nit
    1.00
    rill
    0.99
    daq
    0.98
    Activations Density 0.015%

    No Known Activations