© 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. 20024
    Prev
    Next
    INDEX
    Explanations

    phrases related to undermining or attacking someone's authority

    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
    447
    +0.17
    0.8%
    3
    +0.11
    0.5%
    288
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    3
    +0.17
    0.02
    332
    +0.11
    0.02
    61
    +0.10
    0.02
    Negative Logits
    cise
    -0.74
    area
    -0.73
    lov
    -0.71
    sa
    -0.68
    çĦ
    -0.66
    phe
    -0.65
    lat
    -0.65
    Spoiler
    -0.65
    andro
    -0.64
    alone
    -0.62
    POSITIVE LOGITS
     morale
    0.93
     confidence
    0.87
     livelihood
    0.86
     havoc
    0.86
     perceptions
    0.85
     credibility
    0.83
     democracy
    0.78
     democratically
    0.78
     defences
    0.77
     democracies
    0.77
    Activations Density 0.051%

    No Known Activations