© 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. 9307
    Prev
    Next
    INDEX
    Explanations

    mentions of legal consent in different contexts

    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
    508
    +0.15
    0.7%
    288
    +0.12
    0.6%
    339
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    508
    +0.15
    0.01
    214
    +0.12
    0.01
    575
    +0.11
    0.01
    Negative Logits
     Tycoon
    -0.83
    vae
    -0.72
    nas
    -0.71
    Lens
    -0.68
    GPU
    -0.68
    mania
    -0.68
    gaard
    -0.68
    than
    -0.63
    nikov
    -0.62
    tin
    -0.62
    POSITIVE LOGITS
     decree
    1.07
     consent
    0.99
    ing
    0.88
     decre
    0.88
    ignt
    0.81
     Consent
    0.80
    ible
    0.79
    itional
    0.77
     authorizing
    0.76
    ibility
    0.74
    Activations Density 0.015%

    No Known Activations