© 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. 16624
    Prev
    Next
    INDEX
    Explanations

    pronouns or possessive pronouns related to missing persons

    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
    112
    +0.12
    0.6%
    362
    +0.12
    0.5%
    15
    +0.12
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    15
    +0.12
    0.09
    362
    +0.12
    0.07
    694
    +0.12
    0.04
    Negative Logits
    liga
    -0.88
    deals
    -0.88
    ãĥ´ãĤ¡
    -0.80
    zan
    -0.79
    leans
    -0.76
    nir
    -0.76
    Allows
    -0.73
    Games
    -0.73
    atoon
    -0.72
    reddits
    -0.72
    POSITIVE LOGITS
     whereabouts
    1.45
     belongings
    1.31
     disappearance
    1.22
     plight
    1.20
     assailant
    1.11
     assailants
    1.10
     own
    1.06
     fingerprints
    1.04
     relatives
    1.04
     parents
    1.03
    Activations Density 0.297%

    No Known Activations