© 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. 1912
    Prev
    Next
    INDEX
    Explanations

    information related to safety precautions or recommendations

    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
    640
    +0.23
    1.1%
    497
    +0.13
    0.6%
    496
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    640
    +0.23
    0.15
    6
    +0.13
    0.12
    107
    +0.10
    0.08
    Negative Logits
     triumphant
    -0.82
     sovereignty
    -0.79
    iannopoulos
    -0.76
     collaborator
    -0.76
     presided
    -0.73
     awaited
    -0.72
     inco
    -0.72
     tortured
    -0.71
     imprisoned
    -0.70
     unbeaten
    -0.70
    POSITIVE LOGITS
    Types
    1.34
     Depending
    1.32
    Avoid
    1.26
    Generally
    1.24
    Depending
    1.20
     Examples
    1.20
     Symptoms
    1.19
     Often
    1.18
     Avoid
    1.17
     Tips
    1.17
    Activations Density 0.829%

    No Known Activations