© 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. 2577
    Prev
    Next
    INDEX
    Explanations

    phrases indicating preferences, evaluations, or perceptions

    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
    429
    +0.10
    0.5%
    246
    +0.10
    0.5%
    629
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    737
    +0.10
    0.03
    246
    +0.10
    0.03
    389
    +0.09
    0.02
    Negative Logits
    onz
    -0.72
    endars
    -0.66
    Atl
    -0.65
    ãĤ¼
    -0.64
    cul
    -0.63
    ð
    -0.62
     expend
    -0.61
     actionGroup
    -0.61
    ebook
    -0.60
    enter
    -0.60
    POSITIVE LOGITS
     bothers
    0.79
     happened
    0.78
     takeaway
    0.78
     missing
    0.76
     happens
    0.73
     happening
    0.73
     kicker
    0.72
     distinguishing
    0.69
     thing
    0.69
     transpired
    0.68
    Activations Density 0.046%

    No Known Activations