© 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. 15377
    Prev
    Next
    INDEX
    Explanations

    phrases related to threats, consequences, and negotiations

    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
    393
    +0.10
    0.5%
    128
    +0.09
    0.4%
    440
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    484
    +0.10
    0.06
    128
    +0.09
    0.06
    393
    +0.09
    0.06
    Negative Logits
    hered
    -0.70
    watching
    -0.64
    spir
    -0.63
    got
    -0.62
    wow
    -0.61
    ooked
    -0.61
     Yourself
    -0.61
    illin
    -0.60
    thought
    -0.60
    pires
    -0.60
    POSITIVE LOGITS
     hence
    1.22
     consequently
    1.16
     thereby
    1.14
     preferably
    1.08
     possibly
    1.07
     hopefully
    1.04
     thus
    1.00
     someday
    0.99
     eventually
    0.99
     thereafter
    0.98
    Activations Density 0.325%

    No Known Activations