© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    1. Home
    3. Joseph Bloom · Feature Splitting for GPT2-Small
    5. GPT2-Small
    7. Residual Feature Splitting 49152
    9. 8-RES_FS49152-JB
    11. 37366
    Prev
    Next
    INDEX
    Explanations

    phrases related to strategies and political discussions

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs49152-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_49152
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    49,152
    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
    481
    +0.19
    0.9%
    626
    +0.11
    0.5%
    576
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    362
    +0.19
    0.03
    635
    +0.11
    0.02
    626
    +0.10
    0.02
    Negative Logits
    apses
    -0.69
    assador
    -0.67
    uable
    -0.67
    esters
    -0.66
    ocaust
    -0.65
    essee
    -0.64
    agues
    -0.64
    icious
    -0.64
    adelphia
    -0.64
    atoon
    -0.64
    POSITIVE LOGITS
     mantra
    1.28
     stance
    1.27
     principles
    1.26
     tenets
    1.17
     principle
    1.17
     motto
    1.16
     strategy
    1.11
     ideals
    1.07
     stances
    1.06
     regimen
    1.06
    Activations Density 0.146%

    No Known Activations