© 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 EvalsExportsinterp-engineNEWAPI 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. 13844
    Prev
    Next
    INDEX
    Explanations

    phrases related to self-reflection and personal development

    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
    255
    +0.14
    0.6%
    497
    +0.12
    0.5%
    440
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    255
    +0.14
    0.06
    566
    +0.12
    0.06
    756
    +0.11
    -0.01
    Negative Logits
    Discussion
    -0.73
    USS
    -0.73
    osate
    -0.71
    Chair
    -0.65
     Ladies
    -0.65
    luaj
    -0.65
    aft
    -0.65
     Colleges
    -0.64
    ansas
    -0.64
    Pacific
    -0.63
    POSITIVE LOGITS
     automatically
    1.02
     ALWAYS
    1.02
     necessarily
    0.98
     always
    0.95
     stored
    0.93
     usually
    0.89
     instantly
    0.89
     subjective
    0.86
     proportional
    0.83
     metadata
    0.82
    Activations Density 0.587%

    No Known Activations