© 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
    2. Joseph Bloom · Feature Splitting for GPT2-Small
    3. GPT2-Small
    4. Residual Feature Splitting 49152
    5. 8-RES_FS49152-JB
    6. 45743
    Prev
    Next
    INDEX
    Explanations

    terms related to data density and high frequency occurrences

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs49152-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_49152
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    49,152
    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
    232
    +0.12
    0.6%
    208
    +0.09
    0.4%
    493
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    232
    +0.12
    0.01
    284
    +0.09
    0.01
    208
    +0.09
    0.01
    Negative Logits
    è¦ļéĨĴ
    -0.76
    assetsadobe
    -0.74
    æĸ
    -0.74
    à¨
    -0.73
    OHN
    -0.73
     Kirin
    -0.73
     Pru
    -0.71
    GROUND
    -0.69
     behavi
    -0.66
     Rough
    -0.66
    POSITIVE LOGITS
    aminer
    1.22
    actly
    1.18
    posure
    1.11
    cellent
    1.10
    ample
    1.09
    avier
    1.07
    ercise
    1.06
    cellence
    1.06
    odus
    1.06
    haust
    1.02
    Activations Density 0.009%

    No Known Activations