© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsinterp-engineNEWAPI 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. 21054
    Prev
    Next
    INDEX
    Explanations

    personal pronouns (particularly "me") followed by an emphasis on trust or disbelief

    oai_token-act-pair · gpt-3.5-turboTriggered 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
    373
    +0.14
    0.6%
    88
    +0.11
    0.5%
    511
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    377
    +0.14
    0.03
    391
    +0.11
    0.02
    212
    +0.10
    0.03
    Negative Logits
    affiliated
    -0.77
    iliated
    -0.73
    selage
    -0.71
    ourses
    -0.70
    rane
    -0.66
    ithub
    -0.64
    ibaba
    -0.64
    owship
    -0.64
    kinson
    -0.63
    cific
    -0.62
    POSITIVE LOGITS
     kidding
    0.72
     Charlotte
    0.67
     sarc
    0.67
    !:
    0.67
    =]
    0.63
    ingo
    0.60
     Camel
    0.59
     bro
    0.58
     Galile
    0.57
     listeners
    0.57
    Activations Density 0.136%

    No Known Activations