© 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 EvalsExportsAPI 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. 28035
    Prev
    Next
    INDEX
    Explanations

    the word "though" used in various contexts

    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
    478
    +0.12
    0.7%
    36
    +0.12
    0.7%
    494
    +0.11
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    36
    +0.12
    0.04
    480
    +0.12
    0.04
    366
    +0.11
    0.04
    Negative Logits
    ented
    -2.04
    opsis
    -1.60
    eller
    -1.57
    ents
    -1.56
     forms
    -1.55
     form
    -1.54
     encode
    -1.54
    atories
    -1.52
    aurus
    -1.50
     comprises
    -1.50
    POSITIVE LOGITS
    fully
    1.65
    wolf
    1.64
    ĻĤ
    1.60
    ily
    1.53
    xit
    1.47
    liqu
    1.44
    lessly
    1.42
    mort
    1.42
    ¤
    1.42
     warnings
    1.40
    Activations Density 0.138%

    No Known Activations