© 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 49152
    9. 8-RES_FS49152-JB
    11. 26582
    Prev
    Next
    INDEX
    Explanations

    references to dates and times

    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
    373
    +0.12
    0.6%
    569
    +0.11
    0.5%
    602
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    631
    +0.12
    0.02
    223
    +0.11
    0.02
    525
    +0.10
    0.02
    Negative Logits
     Oblivion
    -0.70
     Furious
    -0.68
     Keeper
    -0.62
     gestures
    -0.60
     Dial
    -0.59
    Reviewer
    -0.57
     controls
    -0.57
     calls
    -0.56
    adder
    -0.55
     resemb
    -0.54
    POSITIVE LOGITS
    cember
    1.44
    ruary
    1.34
    eteenth
    1.07
    EMBER
    1.07
    uary
    1.05
    Feb
    1.04
    ember
    1.02
     06
    0.99
     2018
    0.99
    nesday
    0.99
    Activations Density 0.033%

    No Known Activations