© 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. 4-RES-SM
    11. 14466
    Prev
    Next
    INDEX
    Explanations

    sections or references in structured formats, like reports or academic papers

    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/4-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.4.hook_resid_post
    Hook Layer
    4
    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
    271
    +0.16
    0.9%
    355
    +0.14
    0.8%
    255
    +0.12
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    3
    +0.16
    0.00
    289
    +0.14
    0.00
    506
    +0.12
    0.00
    Negative Logits
     whole
    -1.65
    iginally
    -1.64
     entire
    -1.57
    atically
    -1.53
    ning
    -1.47
     membership
    -1.43
     maker
    -1.42
     builder
    -1.41
     resulting
    -1.41
    ket
    -1.39
    POSITIVE LOGITS
    ĻĤ
    4.37
    Ļª
    4.32
    £
    3.91
    ı
    3.70
    ´
    3.61
    ¼
    3.60
    Ģ
    3.58
    ·
    3.57
    ŀ
    3.52
    Ĵ
    3.46
    Activations Density 0.017%

    No Known Activations