© 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. 1-RES-SM
    11. 25341
    Prev
    Next
    INDEX
    Explanations

    phrases related to numerical representation or frequency within a context

    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/1-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.1.hook_resid_post
    Hook Layer
    1
    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
    12
    +0.13
    0.7%
    54
    +0.10
    0.6%
    23
    +0.10
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    12
    +0.13
    0.03
    374
    +0.10
    0.03
    299
    +0.10
    0.02
    Negative Logits
    ¾
    -2.87
    ¿½
    -2.68
    ¬
    -2.60
    Ń
    -2.59
    ´
    -2.55
    °
    -2.52
    ®
    -2.52
    Ħ
    -2.51
    ¿
    -2.49
    į
    -2.46
    POSITIVE LOGITS
    innings
    1.78
    alion
    1.69
     trouble
    1.45
     south
    1.43
     Highway
    1.43
    versed
    1.41
    uries
    1.37
    avez
    1.34
    pine
    1.34
     nights
    1.31
    Activations Density 0.329%

    No Known Activations