© 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. 156
    Prev
    Next
    INDEX
    Explanations

    mentions of actions or characteristics that are considered excessive or extreme

    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
    289
    +0.11
    0.5%
    566
    +0.11
    0.5%
    85
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    151
    +0.11
    0.03
    763
    +0.11
    0.02
    285
    +0.10
    0.02
    Negative Logits
    ector
    -0.88
    rooms
    -0.83
    stones
    -0.83
    imore
    -0.81
    stone
    -0.80
    iens
    -0.79
    DragonMagazine
    -0.78
    ect
    -0.77
    ians
    -0.76
    ois
    -0.73
    POSITIVE LOGITS
     reliance
    0.99
     burdens
    0.97
     burden
    0.94
     simplistic
    0.87
     indul
    0.86
     infl
    0.84
     aggressive
    0.82
     excessive
    0.82
     amounts
    0.81
     len
    0.81
    Activations Density 0.053%

    No Known Activations