© 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. Under Peer Review · Sparse Autoencoders for Pythia-70M-Deduped
    5. Pythia-70M-Deduped
    7. Residuals
    9. 1-RES-SM
    11. 25012
    Prev
    Next
    INDEX
    Explanations

    greetings and variations of "Hello."

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    ctigges/pythia-70m-deduped__res-sm_processed/1-res-sm
    Prompts (Dashboard)
    32,768 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    32,768
    Data Type
    torch.float32
    Hook Name
    blocks.1.hook_resid_post
    Hook Layer
    1
    Architecture
    standard
    Context Size
    128
    Dataset
    EleutherAI/the_pile_deduplicated
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Neuron Alignment
    Index
    Value
    % of L₁
    156
    +0.17
    1.0%
    410
    +0.13
    0.7%
    365
    +0.13
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    334
    +0.17
    0.02
    156
    +0.13
    0.02
    273
    +0.13
    0.01
    Negative Logits
    ·¸
    -4.04
    ŀ
    -3.94
    ´
    -3.93
    ¸
    -3.91
    Ĺ
    -3.87
    į
    -3.86
    ı
    -3.82
    ĵ
    -3.80
    ³
    -3.79
    IJ
    -3.79
    POSITIVE LOGITS
     messenger
    1.74
     World
    1.72
    starter
    1.71
    eting
    1.66
     unity
    1.52
     Kitty
    1.50
     friends
    1.48
     hello
    1.44
    joy
    1.44
     prayers
    1.43
    Activations Density 0.162%

    No Known Activations