© 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 EvalsExportsAPI 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. 42869
    Prev
    Next
    INDEX
    Explanations

    instances of the word "targeted" and its variations, indicating discussions of directed actions or interventions

    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
    497
    +0.14
    0.7%
    288
    +0.11
    0.5%
    743
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    743
    +0.14
    0.01
    686
    +0.11
    0.01
    710
    +0.11
    0.01
    Negative Logits
    fo
    -0.79
    iHUD
    -0.76
    cess
    -0.74
    loo
    -0.71
    osit
    -0.69
    john
    -0.68
    Struct
    -0.68
     largeDownload
    -0.68
    aug
    -0.67
    IGH
    -0.67
    POSITIVE LOGITS
     killings
    0.96
     targeting
    0.91
     targeted
    0.90
     assass
    0.89
     squarely
    0.86
    ted
    0.84
     targets
    0.80
     towards
    0.79
     Parenthood
    0.78
    ãĤ°
    0.77
    Activations Density 0.010%

    No Known Activations