© 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 12288
    9. 8-RES_FS12288-JB
    11. 10509
    Prev
    Next
    INDEX
    Explanations

    terms related to tax policies and regulations, especially regarding imports, deductions, and exemptions

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs12288-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_12288
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    12,288
    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
    570
    +0.11
    0.5%
    600
    +0.11
    0.5%
    288
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    288
    +0.11
    0.12
    67
    +0.11
    0.08
    600
    +0.10
    0.06
    Negative Logits
    canon
    -0.73
    ortium
    -0.73
     Canaver
    -0.71
    htaking
    -0.70
    agame
    -0.70
    urat
    -0.70
    itcher
    -0.69
    undrum
    -0.68
     Jury
    -0.68
    ĸļ
    -0.68
    POSITIVE LOGITS
     activities
    1.10
     goods
    1.06
     products
    1.05
     components
    1.04
     stimuli
    1.03
     objects
    1.02
     goodies
    1.02
     assets
    1.01
     materials
    1.01
     entities
    1.00
    Activations Density 0.787%

    No Known Activations