© 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. 5819
    Prev
    Next
    INDEX
    Explanations

    words related to environmental disasters and conservation efforts

    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
    575
    +0.16
    0.7%
    288
    +0.13
    0.6%
    547
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    575
    +0.16
    0.02
    288
    +0.13
    0.02
    619
    +0.11
    0.02
    Negative Logits
    reme
    -0.65
    equal
    -0.63
    WithNo
    -0.63
    eq
    -0.63
    erva
    -0.62
    heartedly
    -0.61
    numbered
    -0.61
    CHA
    -0.59
    nos
    -0.59
    bara
    -0.58
    POSITIVE LOGITS
    atory
    0.79
    ring
    0.72
    ongo
    0.69
    aux
    0.68
    ozo
    0.67
     migrating
    0.65
    ocity
    0.65
     migr
    0.64
    owicz
    0.64
    ugg
    0.63
    Activations Density 0.032%

    No Known Activations