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

    Neuronpedia

    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. 256
    Prev
    Next
    INDEX
    Explanations

    phrases related to toxic substances and contamination

    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
    101
    +0.10
    0.5%
    281
    +0.09
    0.4%
    93
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    93
    +0.10
    0.01
    444
    +0.09
    0.01
    99
    +0.09
    0.01
    Negative Logits
    soDeliveryDate
    -1.11
    nee
    -0.82
    things
    -0.80
     PsyNet
    -0.79
    ãĥ¯ãĥ³
    -0.71
    neys
    -0.69
    ãģ¦
    -0.68
    istani
    -0.67
    à©
    -0.67
    abases
    -0.67
    POSITIVE LOGITS
     tremend
    1.28
     chloride
    1.00
     oxide
    0.89
     trem
    0.81
     arsen
    0.78
    osate
    0.78
     dioxide
    0.78
     magnets
    0.71
     titan
    0.71
    s
    0.69
    Activations Density 0.028%

    No Known Activations