© 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 EvalsExportsAPI 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. 19747
    Prev
    Next
    INDEX
    Explanations

    politically charged keywords and names, with a focus on specific individuals and terms related to governmental and political activities

    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
    297
    +0.09
    0.4%
    503
    +0.09
    0.4%
    425
    +0.08
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    232
    +0.09
    0.06
    503
    +0.09
    0.05
    251
    +0.08
    0.05
    Negative Logits
    hyde
    -0.77
     Annotations
    -0.70
     dwarves
    -0.69
     Fargo
    -0.68
     substitutes
    -0.67
     learners
    -0.65
     Rhod
    -0.64
     constructs
    -0.64
     portals
    -0.62
     extracts
    -0.62
    POSITIVE LOGITS
    agging
    1.07
    acious
    1.06
    ithering
    1.04
    assion
    1.03
    anky
    1.02
    ounding
    1.02
    uper
    1.00
    angu
    0.98
    attering
    0.98
    isive
    0.98
    Activations Density 0.220%

    No Known Activations