© 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. 7046
    Prev
    Next
    INDEX
    Explanations

    terms related to software installation and configuration, specifically focusing on RPM files and caching mechanisms

    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
    564
    +0.11
    0.5%
    570
    +0.11
    0.5%
    378
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    288
    +0.11
    0.08
    564
    +0.11
    0.07
    409
    +0.10
    0.05
    Negative Logits
    nos
    -0.75
    hon
    -0.74
    hig
    -0.72
    rupulous
    -0.70
    LOS
    -0.67
     Freem
    -0.66
    othe
    -0.66
    nen
    -0.65
    Weather
    -0.65
    Minecraft
    -0.64
    POSITIVE LOGITS
     molecule
    0.88
     holdings
    0.82
     configuration
    0.81
     activity
    0.80
    acters
    0.78
     pointer
    0.78
     system
    0.78
     stimuli
    0.77
     identifier
    0.76
     output
    0.76
    Activations Density 0.380%

    No Known Activations