© 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 3072
    9. 8-RES_FS3072-JB
    11. 786
    Prev
    Next
    INDEX
    Explanations

    proper nouns, particularly names of 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_fs3072-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_3072
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    3,072
    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
    152
    +0.14
    0.6%
    339
    +0.14
    0.6%
    696
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    696
    +0.14
    0.06
    152
    +0.14
    0.06
    52
    +0.10
    0.05
    Negative Logits
    âĹ¼
    -0.76
    ħĭ
    -0.68
    anwhile
    -0.68
    udic
    -0.67
    llor
    -0.65
    ï¸ı
    -0.65
    PDATE
    -0.65
    rule
    -0.63
    ļé
    -0.63
    phal
    -0.62
    POSITIVE LOGITS
    't
    1.12
    nie
    1.06
     Draper
    1.05
    atives
    0.95
    ations
    0.93
    atell
    0.93
    ovan
    0.91
    ners
    0.87
    ny
    0.86
    ative
    0.85
    Activations Density 0.438%

    No Known Activations