© 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. 182
    Prev
    Next
    INDEX
    Explanations

    mentions of e-cigarettes and related topics such as bans, regulations, public use, government involvement, and survey results

    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
    447
    +0.20
    0.9%
    481
    +0.13
    0.6%
    708
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    288
    +0.20
    0.05
    343
    +0.13
    0.04
    181
    +0.10
    0.04
    Negative Logits
     Seym
    -0.78
     Ful
    -0.69
     Mutual
    -0.69
     Minotaur
    -0.68
     Ow
    -0.67
     Icar
    -0.66
     Niet
    -0.66
     Ambrose
    -0.65
     Schw
    -0.65
     Fernand
    -0.64
    POSITIVE LOGITS
    ervative
    0.90
    ervatives
    0.80
     supporters
    0.78
    âĦ¢
    0.74
     proponents
    0.70
    OTA
    0.67
    ONSORED
    0.67
     campaigners
    0.65
     adherents
    0.64
     aka
    0.63
    Activations Density 0.163%

    No Known Activations