© 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. 16008
    Prev
    Next
    INDEX
    Explanations

    references to geographic locations and events happening in those areas, particularly related to safety concerns

    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
    362
    +0.18
    0.8%
    393
    +0.18
    0.8%
    317
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    362
    +0.18
    0.11
    393
    +0.18
    0.07
    636
    +0.09
    0.06
    Negative Logits
    thood
    -0.87
    pelling
    -0.84
    TION
    -0.80
    uci
    -0.79
    isson
    -0.77
    utan
    -0.76
    HCR
    -0.72
    heed
    -0.70
    maid
    -0.70
    olkien
    -0.69
    POSITIVE LOGITS
     halls
    1.23
     periphery
    1.23
     globe
    1.22
     streets
    1.22
     skies
    1.15
     shores
    1.12
     corridors
    1.12
     entirety
    1.10
     entire
    1.07
     nation
    1.06
    Activations Density 0.340%

    No Known Activations