© 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. 22672
    Prev
    Next
    INDEX
    Explanations

    words related to ascent or rising, particularly in a metaphorical or symbolic context

    oai_token-act-pair · gpt-4o-miniTriggered 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
    365
    +0.12
    0.6%
    311
    +0.11
    0.5%
    300
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    503
    +0.12
    0.01
    749
    +0.11
    0.01
    365
    +0.10
    0.01
    Negative Logits
     loose
    -0.65
    tery
    -0.65
    chy
    -0.64
    meal
    -0.63
     Bastard
    -0.62
     Wonderland
    -0.62
     backdoor
    -0.61
     Dragonbound
    -0.61
     dearly
    -0.61
     gallery
    -0.61
    POSITIVE LOGITS
    ribed
    1.67
    ribing
    1.66
    endant
    1.44
    ribe
    1.42
    ript
    1.41
    ension
    1.32
    ribes
    1.29
    ended
    1.27
    endants
    1.25
    ensible
    1.24
    Activations Density 0.024%

    No Known Activations