© 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. 1910
    Prev
    Next
    INDEX
    Explanations

    mentions of human rights-related terms and legal actions

    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
    515
    +0.12
    0.6%
    332
    +0.12
    0.5%
    288
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    332
    +0.12
    0.05
    515
    +0.12
    0.04
    719
    +0.10
    0.04
    Negative Logits
    forth
    -0.73
     Transcript
    -0.72
    onne
    -0.69
    creen
    -0.68
    urations
    -0.67
    RET
    -0.65
    eryl
    -0.64
    etsy
    -0.63
    OHN
    -0.61
     Referred
    -0.60
    POSITIVE LOGITS
    itarian
    1.31
     beings
    1.31
    itar
    1.26
    istic
    1.11
    itary
    1.03
    oids
    1.03
    izing
    1.02
     rights
    1.01
    izes
    0.97
    ization
    0.95
    Activations Density 0.051%

    No Known Activations