© 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 768
    9. 8-RES_FS768-JB
    11. 640
    Prev
    Next
    INDEX
    Explanations

    references to water-related terms and places

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs768-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_768
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    768
    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
    592
    +0.11
    0.5%
    459
    +0.10
    0.5%
    322
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    288
    +0.11
    0.21
    232
    +0.10
    0.12
    419
    +0.10
    0.11
    Negative Logits
    eering
    -0.99
    Reloaded
    -0.86
     tremend
    -0.69
    CVE
    -0.65
    -+-+-+-+
    -0.64
    ures
    -0.63
    iven
    -0.63
    gnu
    -0.63
    ijah
    -0.61
    ignment
    -0.60
    POSITIVE LOGITS
    melon
    1.21
     basin
    1.06
    colour
    1.00
    water
    0.99
    maid
    0.95
    front
    0.94
     baptism
    0.92
    Bott
    0.91
     ashore
    0.89
    tub
    0.86
    Activations Density 4.973%

    No Known Activations