© 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. 2-RES-SM
    11. 12661
    Prev
    Next
    INDEX
    Explanations

    instances of the word "mis" and its variations, indicating themes of misunderstanding or misuse

    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
    338
    +0.15
    0.8%
    59
    +0.14
    0.8%
    344
    +0.13
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    322
    +0.15
    0.01
    344
    +0.14
    0.03
    134
    +0.13
    0.02
    Negative Logits
    ogether
    -1.75
    cgi
    -1.52
    AtIndex
    -1.50
    à°¿
    -1.49
     matter
    -1.46
    markdown
    -1.41
    *](#
    -1.41
     seja
    -1.41
     absence
    -1.40
     later
    -1.38
    POSITIVE LOGITS
    ¿
    2.92
    Ħ
    2.87
    ĩ
    2.83
    Ń
    2.77
    ª
    2.70
    ·
    2.69
    ¹
    2.63
    ¨
    2.63
    µ
    2.59
    §
    2.52
    Activations Density 0.165%

    No Known Activations