© 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. 3208
    Prev
    Next
    INDEX
    Explanations

    phrases related to downfall or negative outcomes for individuals or entities

    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
    151
    +0.13
    0.6%
    373
    +0.10
    0.5%
    523
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    151
    +0.13
    0.03
    9
    +0.10
    0.02
    113
    +0.10
    0.02
    Negative Logits
    ¶ħ
    -0.82
    ilies
    -0.73
    iliary
    -0.72
    ificent
    -0.69
    ivity
    -0.67
    ago
    -0.64
    rehens
    -0.63
    Us
    -0.63
    ially
    -0.63
    cius
    -0.62
    POSITIVE LOGITS
     stairs
    1.14
    graded
    0.97
    grading
    0.93
    stairs
    0.90
    hill
    0.88
     horm
    0.87
    river
    0.87
    LOAD
    0.85
     hill
    0.82
    pour
    0.80
    Activations Density 0.029%

    No Known Activations