© 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 24576
    9. 8-RES_FS24576-JB
    11. 19337
    Prev
    Next
    INDEX
    Explanations

    phrases related to moral or ethical judgments

    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
    253
    +0.14
    0.6%
    288
    +0.10
    0.5%
    159
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    253
    +0.14
    0.02
    242
    +0.10
    0.02
    151
    +0.09
    0.02
    Negative Logits
    ilant
    -0.88
    craft
    -0.80
    lets
    -0.77
    avers
    -0.76
    oling
    -0.75
    ocket
    -0.74
    cest
    -0.73
    planes
    -0.72
    yss
    -0.71
    frey
    -0.71
    POSITIVE LOGITS
     deviations
    0.86
     behaviour
    0.80
     Danger
    0.78
     behavior
    0.77
    ible
    0.75
     compromises
    0.74
     standards
    0.74
     norms
    0.73
     srfAttach
    0.72
     acceptable
    0.71
    Activations Density 0.047%

    No Known Activations