© 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 12288
    9. 8-RES_FS12288-JB
    11. 1726
    Prev
    Next
    INDEX
    Explanations

    negative or alarming information related to government actions and consequences

    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
    373
    +0.17
    0.8%
    231
    +0.10
    0.5%
    483
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    373
    +0.17
    -0.04
    671
    +0.10
    0.05
    348
    +0.10
    0.07
    Negative Logits
    undy
    -0.58
    SHIP
    -0.57
    iste
    -0.56
    #$
    -0.56
    ngth
    -0.55
    dinand
    -0.54
    debian
    -0.53
    cientious
    -0.52
    ordered
    -0.52
    hyde
    -0.51
    POSITIVE LOGITS
     importantly
    0.90
     aside
    0.84
     furthermore
    0.82
     interestingly
    0.78
     moreover
    0.75
    versely
    0.74
     notwithstanding
    0.73
     meantime
    0.68
     meanwhile
    0.67
    identally
    0.66
    Activations Density 0.811%

    No Known Activations