© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it
    Neuronpedia
    Jacobian LensNEW
    Natural Language
    Autoencoders
    MAY 26    Assistant AxisJAN 26Circuit TracerAPR 26SteerSAE EvalsExportsinterp-engineNEW 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. 3079
    Prev
    Next
    INDEX
    Explanations

    strong emotional words or phrases with a focus on specific nouns like 'victim' or 'it'

    oai_token-act-pair · gpt-3.5-turbo
    Triggered 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
    Neuron Alignment
    Index
    Value
    % of L₁
    272
    +0.10
    0.4%
    567
    +0.09
    0.4%
    35
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    567
    +0.10
    0.03
    431
    +0.09
    0.03
    213
    +0.09
    0.02
    Negative Logits
    hement
    -0.82
    imaru
    -0.80
    vertisement
    -0.76
    igham
    -0.75
    ģĸ
    -0.74
    swick
    -0.70
    ONSORED
    -0.68
    rative
    -0.66
    erate
    -0.64
    ittal
    -0.64
    POSITIVE LOGITS
     whereabouts
    0.85
     drill
    0.79
     firsthand
    0.75
     intimately
    0.73
     instinctively
    0.72
    âĨij
    0.71
    CHAT
    0.67
     exists
    0.67
     secret
    0.66
     secrets
    0.66
    Activations Density 0.149%

    No Known Activations