© 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. 23965
    Prev
    Next
    INDEX
    Explanations

    negative phrases and assertions

    oai_token-act-pair · gpt-4o-miniTriggered 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
    241
    +0.10
    0.5%
    36
    +0.10
    0.4%
    71
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    616
    +0.10
    0.04
    376
    +0.10
    0.04
    754
    +0.09
    0.04
    Negative Logits
    iership
    -0.69
    RAFT
    -0.66
    igan
    -0.64
    imbabwe
    -0.64
    run
    -0.63
    ËĪ
    -0.63
    ospons
    -0.63
    abal
    -0.62
    MpServer
    -0.62
    gru
    -0.61
    POSITIVE LOGITS
     kidding
    1.20
     doubt
    1.04
     matter
    1.03
     wonder
    1.00
     longer
    0.96
    xious
    0.95
    etheless
    0.92
     excuses
    0.90
    zzle
    0.85
     sooner
    0.85
    Activations Density 0.052%

    No Known Activations