© 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. 14490
    Prev
    Next
    INDEX
    Explanations

    terms related to gaining influence or taking hold of something

    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
    447
    +0.31
    1.5%
    225
    +0.10
    0.5%
    181
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    225
    +0.31
    0.05
    724
    +0.10
    0.06
    665
    +0.10
    0.05
    Negative Logits
     countdown
    -0.71
     Declaration
    -0.61
     condol
    -0.57
     cardinal
    -0.56
     constant
    -0.56
     lockout
    -0.56
     blackout
    -0.55
     vomit
    -0.55
     severity
    -0.55
     istg
    -0.55
    POSITIVE LOGITS
     overseas
    1.06
     elsewhere
    1.04
     abroad
    0.99
     internationally
    0.92
     nationally
    0.89
     worldwide
    0.88
     domestically
    0.86
    hare
    0.86
     nationwide
    0.81
     globally
    0.78
    Activations Density 0.292%

    No Known Activations