© 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. 11844
    Prev
    Next
    INDEX
    Explanations

    names of specific people

    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
    373
    +0.13
    0.6%
    223
    +0.11
    0.5%
    311
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    373
    +0.13
    -0.00
    232
    +0.11
    0.01
    99
    +0.11
    0.01
    Negative Logits
     ashore
    -0.71
     Siren
    -0.65
     Clash
    -0.62
    dule
    -0.62
    izational
    -0.60
    abeth
    -0.57
     Dragonbound
    -0.57
     Carmen
    -0.57
    verson
    -0.56
     braces
    -0.56
    POSITIVE LOGITS
    eatures
    1.05
    ortun
    1.02
    lex
    0.97
    unction
    0.97
    ornia
    0.95
    lect
    0.95
    req
    0.93
    ruit
    0.90
    icient
    0.88
    rame
    0.88
    Activations Density 0.014%

    No Known Activations