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

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports 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. 19843
    Prev
    Next
    INDEX
    Explanations

    phrases related to cats

    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
    167
    +0.14
    0.6%
    361
    +0.10
    0.5%
    289
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    361
    +0.14
    0.02
    167
    +0.10
    0.01
    289
    +0.10
    0.01
    Negative Logits
    mble
    -0.76
    hower
    -0.76
     htt
    -0.68
     Seym
    -0.68
    Reloaded
    -0.67
    ij士
    -0.66
     Vaugh
    -0.63
     Fellowship
    -0.62
    eous
    -0.61
    assetsadobe
    -0.61
    POSITIVE LOGITS
    aclysm
    1.66
    heter
    1.35
    apult
    1.31
    cher
    1.17
    chers
    1.15
    alogue
    1.13
    hedral
    1.11
    alyst
    1.04
    fish
    1.03
    alog
    1.02
    Activations Density 0.032%

    No Known Activations