© 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. 22637
    Prev
    Next
    INDEX
    Explanations

    phrases that express notions of deservingness and recognition

    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
    466
    +0.11
    0.5%
    719
    +0.10
    0.5%
    696
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    533
    +0.11
    0.02
    608
    +0.10
    0.03
    466
    +0.09
    0.02
    Negative Logits
     renov
    -0.67
    ynchron
    -0.66
    ullivan
    -0.65
     synchronized
    -0.65
     portals
    -0.64
     inefficient
    -0.62
    nels
    -0.61
    ashi
    -0.60
     improvised
    -0.58
    iry
    -0.58
    POSITIVE LOGITS
     THANK
    0.94
     praise
    0.89
     scorn
    0.86
    udos
    0.82
     dignity
    0.81
     privileges
    0.80
    uala
    0.78
     rightfully
    0.77
     ridicule
    0.76
     admiration
    0.75
    Activations Density 0.116%

    No Known Activations