© 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. 31765
    Prev
    Next
    INDEX
    Explanations

    instances of quotation marks and their associated content

    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
    102
    +0.14
    0.8%
    47
    +0.13
    0.7%
    103
    +0.13
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    103
    +0.14
    0.04
    102
    +0.13
    0.04
    124
    +0.13
    0.04
    Negative Logits
     submissions
    -1.58
     guilty
    -1.56
     ‘
    -1.55
     footage
    -1.53
    geries
    -1.46
    .[]{
    -1.44
    .[^
    -1.42
     lately
    -1.38
    mitted
    -1.37
     athletes
    -1.37
    POSITIVE LOGITS
    į
    2.88
    ĭ
    2.57
    ³
    2.53
    ¸
    2.44
    ¢
    2.39
    IJ
    2.24
    ¿
    2.21
    ľĵ
    2.21
    ¬
    2.18
    Ķ
    2.16
    Activations Density 0.129%

    No Known Activations