© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    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 49152
    9. 8-RES_FS49152-JB
    11. 24018
    Prev
    Next
    INDEX
    Explanations

    various forms of the verb "to rate" and related terms

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs49152-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_49152
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    49,152
    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
    81
    +0.11
    0.5%
    566
    +0.10
    0.5%
    28
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    288
    +0.11
    0.03
    81
    +0.10
    0.03
    660
    +0.10
    0.03
    Negative Logits
    sters
    -0.80
    aldi
    -0.74
    thing
    -0.72
    fully
    -0.71
    older
    -0.70
    adra
    -0.69
    enberg
    -0.68
    while
    -0.68
    stru
    -0.68
    ned
    -0.68
    POSITIVE LOGITS
    eals
    0.75
    aily
    0.72
     BY
    0.72
    irect
    0.70
    ombat
    0.69
     version
    0.68
    ict
    0.66
    escription
    0.65
    urally
    0.65
     medicine
    0.64
    Activations Density 0.084%

    No Known Activations