© 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 24576
    9. 8-RES_FS24576-JB
    11. 23976
    Prev
    Next
    INDEX
    Explanations

    references to violent or aggressive actions

    oai_token-act-pair · gpt-4o-miniTriggered 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
    288
    +0.15
    0.7%
    379
    +0.12
    0.5%
    472
    +0.12
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    601
    +0.15
    0.03
    379
    +0.12
    0.04
    472
    +0.12
    0.03
    Negative Logits
    ãĤ©
    -0.78
    theless
    -0.74
    YC
    -0.67
    zl
    -0.65
    åħī
    -0.63
    inders
    -0.61
    ILLE
    -0.61
    dit
    -0.61
     Jazz
    -0.58
    ETA
    -0.57
    POSITIVE LOGITS
     against
    1.02
    CVE
    0.87
     vector
    0.87
    iveness
    0.82
     attacks
    0.81
     perpetrated
    0.79
    attack
    0.79
     attack
    0.79
    ivist
    0.77
    ivation
    0.77
    Activations Density 0.052%

    No Known Activations