© 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. 8145
    Prev
    Next
    INDEX
    Explanations

    questions or statements in the form of 'how much', 'how many', and 'who would'

    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
    121
    +0.10
    0.4%
    692
    +0.09
    0.4%
    559
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    60
    +0.10
    0.04
    236
    +0.09
    0.04
    263
    +0.09
    0.03
    Negative Logits
    onite
    -0.84
    mud
    -0.82
    Zone
    -0.82
    hetics
    -0.79
    ocre
    -0.78
    thal
    -0.76
    izons
    -0.74
    ĸļ
    -0.74
    Think
    -0.74
    Listen
    -0.72
    POSITIVE LOGITS
     exact
    0.82
     extent
    0.81
     confir
    0.79
     exactly
    0.79
     specifics
    0.78
     prompted
    0.77
     nor
    0.70
     precise
    0.69
     refunds
    0.69
     charges
    0.69
    Activations Density 0.115%

    No Known Activations