© 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. 28483
    Prev
    Next
    INDEX
    Explanations

    instances of the word "to" and related phrases that indicate direction or purpose

    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
    420
    +0.12
    0.7%
    478
    +0.12
    0.6%
    192
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    363
    +0.12
    0.03
    48
    +0.12
    0.03
    420
    +0.10
    0.02
    Negative Logits
    ricular
    -1.53
     sake
    -1.48
    ieux
    -1.44
    ervation
    -1.43
     Reading
    -1.33
    ight
    -1.32
     Judaism
    -1.32
    buntu
    -1.32
    oir
    -1.30
    ism
    -1.29
    POSITIVE LOGITS
    ¡
    1.98
    °
    1.92
    rapeut
    1.90
    ĥ
    1.87
    ½
    1.85
    Ĺ
    1.82
    Ī
    1.78
    Į
    1.77
    £
    1.76
    ¦
    1.75
    Activations Density 0.403%

    No Known Activations