© 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 12288
    9. 8-RES_FS12288-JB
    11. 9617
    Prev
    Next
    INDEX
    Explanations

    textual mentions of official forms or questionnaires

    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
    680
    +0.12
    0.5%
    388
    +0.12
    0.5%
    639
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    680
    +0.12
    0.05
    540
    +0.12
    0.07
    448
    +0.10
    0.03
    Negative Logits
    ~~~~~~~~
    -0.76
    zees
    -0.72
    =]
    -0.71
     gor
    -0.67
    seless
    -0.67
     disrupting
    -0.66
    "]=>
    -0.64
     maniac
    -0.64
     senseless
    -0.64
     antagon
    -0.63
    POSITIVE LOGITS
     paperwork
    1.27
     questionnaire
    1.22
     receipts
    1.18
     ledger
    1.12
     documents
    1.11
     records
    1.10
     invoice
    1.05
     statistics
    1.05
     affidavit
    1.04
     appraisal
    1.04
    Activations Density 0.873%

    No Known Activations