© 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. 13178
    Prev
    Next
    INDEX
    Explanations

    phrases related to legal matters, drug use, law enforcement, and political conversations

    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
    35
    +0.11
    0.5%
    128
    +0.11
    0.5%
    419
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    419
    +0.11
    0.03
    35
    +0.11
    0.03
    461
    +0.11
    0.03
    Negative Logits
    ffen
    -0.74
    auga
    -0.68
    soType
    -0.66
    semble
    -0.65
    #$
    -0.65
    gaard
    -0.65
    hower
    -0.65
    ancies
    -0.64
    vironment
    -0.64
     Finch
    -0.63
    POSITIVE LOGITS
     cartels
    1.15
     trafficking
    1.14
     overdoses
    1.10
     traffickers
    1.10
     overdose
    1.09
    store
    1.07
     cartel
    1.06
     drugs
    1.04
     addicts
    1.01
    stores
    1.00
    Activations Density 0.033%

    No Known Activations