© 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. 9639
    Prev
    Next
    INDEX
    Explanations

    mentions of "Bra" and its variations, likely related to types of bras or undergarments

    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
    28
    +0.14
    0.8%
    504
    +0.12
    0.7%
    453
    +0.11
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    386
    +0.14
    0.01
    355
    +0.12
    0.03
    28
    +0.11
    0.01
    Negative Logits
    ation
    -1.73
    rait
    -1.60
     rest
    -1.48
     tomography
    -1.46
    ry
    -1.44
    ively
    -1.43
    ries
    -1.42
    riv
    -1.42
    vironment
    -1.41
    ologies
    -1.41
    POSITIVE LOGITS
    ¬
    3.57
    į
    3.46
    ĥ½
    3.45
    ½
    3.39
    «
    3.33
    ĻĤ
    3.32
    Ľ
    3.17
    ³
    3.16
    ¿½
    3.16
    Ī
    3.15
    Activations Density 0.179%

    No Known Activations