© 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 EvalsExportsinterp-engineNEWAPI 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. 22897
    Prev
    Next
    INDEX
    Explanations

    phrases that involve making suggestions or requests

    oai_token-act-pair · gpt-4o-miniTriggered 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
    163
    +0.11
    0.5%
    434
    +0.11
    0.5%
    105
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    665
    +0.11
    0.05
    105
    +0.11
    0.04
    49
    +0.11
    0.05
    Negative Logits
    uably
    -0.88
     intrinsically
    -0.83
    quickShipAvailable
    -0.81
    obyl
    -0.80
    etheless
    -0.78
     underestimated
    -0.75
     undeniably
    -0.74
     strikingly
    -0.74
     anecd
    -0.72
     striving
    -0.71
    POSITIVE LOGITS
    —"
    0.90
     escort
    0.89
    opy
    0.88
     join
    0.84
     uh
    0.84
     dinner
    0.82
     lunch
    0.82
     evacuate
    0.82
     please
    0.81
     arrange
    0.80
    Activations Density 0.289%

    No Known Activations