© 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 EvalsExportsinterp-engineNEWAPI 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. 21059
    Prev
    Next
    INDEX
    Explanations

    scientific and academic titles, such as 'scientist', 'researcher', and 'professor', especially with corresponding fields of study

    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
    414
    +0.11
    0.5%
    233
    +0.09
    0.4%
    684
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    204
    +0.11
    0.04
    414
    +0.09
    0.03
    104
    +0.09
    0.03
    Negative Logits
     needles
    -0.79
    ombies
    -0.75
     idiots
    -0.74
    rums
    -0.73
    Players
    -0.71
    Ts
    -0.67
    Reward
    -0.66
    humans
    -0.66
     hordes
    -0.65
    SPONSORED
    -0.65
    POSITIVE LOGITS
     specializing
    1.06
    hunter
    0.85
     adjunct
    0.83
    alyst
    0.81
     assistant
    0.81
     extraord
    0.79
     affiliated
    0.77
    abase
    0.76
     emer
    0.75
     adviser
    0.75
    Activations Density 0.090%

    No Known Activations