© 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 49152
    9. 8-RES_FS49152-JB
    11. 25808
    Prev
    Next
    INDEX
    Explanations

    phrases that argue or advocate for a specific point of view or position

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs49152-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_49152
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    49,152
    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
    642
    +0.10
    0.4%
    276
    +0.09
    0.4%
    447
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    276
    +0.10
    0.01
    324
    +0.09
    0.01
    386
    +0.09
    0.01
    Negative Logits
     Seym
    -0.70
    orks
    -0.66
    liction
    -0.65
    elta
    -0.64
    attery
    -0.63
    ummer
    -0.62
    BW
    -0.62
    leted
    -0.61
    kered
    -0.60
    onder
    -0.60
    POSITIVE LOGITS
     against
    1.06
     convinc
    0.93
    against
    0.88
    Against
    0.86
    cases
    0.77
     Keen
    0.75
     Against
    0.75
     loudly
    0.74
     persu
    0.73
     why
    0.73
    Activations Density 0.021%

    No Known Activations