© 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. 24084
    Prev
    Next
    INDEX
    Explanations

    terms related to organizations, facilities, or systems in a structured or formal context

    oai_token-act-pair · gpt-4o-miniTriggered 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
    224
    +0.11
    0.5%
    288
    +0.11
    0.5%
    736
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    288
    +0.11
    0.07
    620
    +0.11
    0.03
    115
    +0.10
    0.06
    Negative Logits
    pread
    -0.74
    姫
    -0.68
    ricular
    -0.68
    à¨
    -0.68
    ometric
    -0.68
    anish
    -0.66
    å¸
    -0.64
    stroke
    -0.64
    ãĤĮ
    -0.63
    Diff
    -0.63
    POSITIVE LOGITS
     overseen
    0.82
     located
    0.81
     marketplace
    0.71
    eers
    0.69
     website
    0.69
     headquartered
    0.69
     circa
    0.67
     fame
    0.67
     hosted
    0.66
     system
    0.65
    Activations Density 0.326%

    No Known Activations