© 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. 8184
    Prev
    Next
    INDEX
    Explanations

    situations or events described by verbs in an active, present tense

    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
    758
    +0.12
    0.5%
    373
    +0.10
    0.5%
    762
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    418
    +0.12
    0.03
    758
    +0.10
    0.03
    102
    +0.10
    0.03
    Negative Logits
     Stain
    -0.69
     chatter
    -0.66
     Prescott
    -0.62
     Moose
    -0.61
     Dresden
    -0.59
     nap
    -0.59
     sqor
    -0.58
     demographics
    -0.58
     partName
    -0.58
     correlation
    -0.58
    POSITIVE LOGITS
     able
    1.06
    heading
    1.00
     replaced
    0.98
     sure
    0.92
    fitting
    0.89
     considered
    0.88
    fits
    0.88
    ige
    0.84
     judged
    0.83
     seen
    0.83
    Activations Density 0.057%

    No Known Activations