© 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. Under Peer Review · Sparse Autoencoders for Pythia-70M-Deduped
    5. Pythia-70M-Deduped
    7. Residuals
    9. 2-RES-SM
    11. 9924
    Prev
    Next
    INDEX
    Explanations

    greetings and friendly salutations

    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/2-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.2.hook_resid_post
    Hook Layer
    2
    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
    410
    +0.18
    1.0%
    156
    +0.15
    0.8%
    334
    +0.14
    0.8%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    410
    +0.18
    0.03
    334
    +0.15
    0.03
    317
    +0.14
    0.02
    Negative Logits
    ĵ
    -4.04
    ı
    -3.56
    ij
    -3.51
    ħ
    -3.49
    ¯
    -3.49
    ĥ
    -3.46
    IJ
    -3.31
    Ĭ
    -3.30
    ļ
    -3.30
    Ķ
    -3.29
    POSITIVE LOGITS
     friends
    1.94
     Friends
    1.80
     hello
    1.65
    rade
    1.57
     My
    1.54
    atus
    1.50
     Kitty
    1.50
    eting
    1.50
     friend
    1.49
    bered
    1.49
    Activations Density 0.137%

    No Known Activations