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

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    1. Home
    3. Joseph Bloom · Feature Splitting for GPT2-Small
    5. GPT2-Small
    7. Residual Feature Splitting 12288
    9. 8-RES_FS12288-JB
    11. 375
    Prev
    Next
    INDEX
    Explanations

    phrases related to destruction or causing harm to objects or entities

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs12288-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_12288
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    12,288
    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
    +0.14
    0.7%
    375
    +0.14
    0.7%
    269
    +0.12
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    529
    +0.14
    0.03
    375
    +0.14
    0.03
    269
    +0.12
    0.02
    Negative Logits
    DragonMagazine
    -0.76
    soType
    -0.69
    Clar
    -0.69
    REL
    -0.67
    POL
    -0.66
    omez
    -0.65
    Son
    -0.65
     reluct
    -0.65
     Trader
    -0.63
    iatrics
    -0.62
    POSITIVE LOGITS
     havoc
    1.21
    roying
    0.86
    adoes
    0.84
     wrought
    0.80
    ados
    0.78
     piles
    0.75
    IED
    0.75
     wre
    0.72
    ishing
    0.70
     Dresden
    0.69
    Activations Density 0.119%

    No Known Activations