© 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 12288
    9. 8-RES_FS12288-JB
    11. 1284
    Prev
    Next
    INDEX
    Explanations

    verbal actions related to authority or planning

    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
    240
    +0.10
    0.5%
    28
    +0.10
    0.4%
    123
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    326
    +0.10
    0.05
    113
    +0.10
    0.04
    665
    +0.09
    0.05
    Negative Logits
    otive
    -0.67
    ":["
    -0.62
    ions
    -0.61
    able
    -0.60
    ivity
    -0.57
    ment
    -0.57
    holding
    -0.57
     ·
    -0.57
    ministic
    -0.57
    inian
    -0.56
    POSITIVE LOGITS
     up
    1.36
    Up
    1.12
     Up
    1.09
    up
    1.00
    UP
    0.99
     UP
    0.87
    ups
    0.86
     Ups
    0.83
     ups
    0.76
    uled
    0.75
    Activations Density 0.311%

    No Known Activations