© 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 24576
    9. 8-RES_FS24576-JB
    11. 18763
    Prev
    Next
    INDEX
    Explanations

    mentions of political entities, particularly the Kremlin

    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
    288
    +0.15
    0.7%
    526
    +0.13
    0.6%
    481
    +0.12
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    488
    +0.15
    0.01
    200
    +0.13
    0.01
    201
    +0.12
    0.01
    Negative Logits
     Kear
    -0.77
    ORGE
    -0.75
     Hancock
    -0.72
    ayer
    -0.70
     Navajo
    -0.70
    phies
    -0.70
    Thom
    -0.70
    bee
    -0.69
    ascar
    -0.69
    Thompson
    -0.68
    POSITIVE LOGITS
     rall
    0.93
     Kremlin
    0.86
     dictator
    0.84
     dictatorship
    0.83
     annexed
    0.79
     Jinping
    0.78
     sympath
    0.77
     dissidents
    0.74
     meddling
    0.74
     disinformation
    0.74
    Activations Density 0.012%

    No Known Activations