© 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 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. 17937
    Prev
    Next
    INDEX
    Explanations

    terms related to rules, changes, actions, and associations

    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
    480
    +0.25
    1.2%
    447
    +0.22
    1.1%
    627
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    665
    +0.25
    0.08
    608
    +0.22
    0.07
    157
    +0.11
    0.06
    Negative Logits
    die
    -0.75
    å¥
    -0.66
     respectively
    -0.65
     mates
    -0.63
    sein
    -0.61
     ensu
    -0.60
    aten
    -0.59
     comrades
    -0.59
     pist
    -0.58
     teammate
    -0.57
    POSITIVE LOGITS
     FOIA
    0.76
     antitrust
    0.74
     enforcing
    0.73
     licensing
    0.72
     backlog
    0.72
     oversight
    0.71
     outsourcing
    0.69
     Grassley
    0.69
     jurisdiction
    0.69
     hiring
    0.67
    Activations Density 0.862%

    No Known Activations