© 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. 4714
    Prev
    Next
    INDEX
    Explanations

    phrases related to standing by and defending different entities, specifically in a legal or political context

    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
    78
    +0.10
    0.5%
    3
    +0.10
    0.4%
    8
    +0.08
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    78
    +0.10
    0.01
    587
    +0.10
    0.02
    435
    +0.08
    0.01
    Negative Logits
    ãĤ¼ãĤ¦ãĤ¹
    -0.69
    uscript
    -0.68
    Course
    -0.67
    igon
    -0.66
    owler
    -0.66
     trave
    -0.65
     proble
    -0.65
    ilib
    -0.64
    Lear
    -0.64
    elf
    -0.64
    POSITIVE LOGITS
     steadfast
    0.91
     sidelines
    0.91
     vigil
    0.80
     Defenders
    0.76
     uncond
    0.74
     indef
    0.72
    ardless
    0.69
     principles
    0.69
     Against
    0.69
     defenders
    0.68
    Activations Density 0.057%

    No Known Activations