© 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 12288
    9. 8-RES_FS12288-JB
    11. 10802
    Prev
    Next
    INDEX
    Explanations

    words related to admiration and strong emotional connections towards sports figures

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs12288-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_12288
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    12,288
    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
    373
    +0.11
    0.5%
    431
    +0.09
    0.4%
    467
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    351
    +0.11
    0.04
    285
    +0.09
    0.03
    96
    +0.09
    0.04
    Negative Logits
     transpired
    -0.71
    utor
    -0.69
     deposition
    -0.67
    uto
    -0.64
     preliminary
    -0.64
    Correct
    -0.64
    visible
    -0.63
    imester
    -0.63
     transcript
    -0.63
    Dispatch
    -0.63
    POSITIVE LOGITS
     dearly
    1.14
     hates
    1.08
     uncond
    1.07
     cherish
    1.04
    hate
    1.00
     adore
    0.99
     passionately
    0.98
     loves
    0.97
     despise
    0.94
     hating
    0.93
    Activations Density 0.746%

    No Known Activations