© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsinterp-engineNEWAPI 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. 5577
    Prev
    Next
    INDEX
    Explanations

    terms related to caching, particularly cache entries 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
    288
    +0.18
    0.8%
    442
    +0.13
    0.6%
    447
    +0.12
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    288
    +0.18
    0.02
    326
    +0.13
    0.01
    442
    +0.12
    0.01
    Negative Logits
    lihood
    -0.74
    sky
    -0.69
    inse
    -0.69
     Cambod
    -0.67
    borough
    -0.66
    hani
    -0.66
    stad
    -0.65
    abouts
    -0.64
     Narc
    -0.62
     Asuka
    -0.62
    POSITIVE LOGITS
    xia
    1.06
     cache
    0.94
     dumps
    0.88
     Cache
    0.86
     caches
    0.80
    cache
    0.74
     cached
    0.74
     storage
    0.73
     caching
    0.73
     layer
    0.72
    Activations Density 0.021%

    No Known Activations