© 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. 27118
    Prev
    Next
    INDEX
    Explanations

    specific references to technical terminology and identifiers

    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
    369
    +0.20
    1.1%
    489
    +0.13
    0.7%
    235
    +0.12
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    369
    +0.20
    0.04
    143
    +0.13
    0.06
    112
    +0.12
    0.06
    Negative Logits
    uity
    -1.64
    iency
    -1.63
     '{
    -1.59
    uities
    -1.57
     flavors
    -1.46
    osition
    -1.44
    .*,
    -1.43
    overs
    -1.42
     ‘
    -1.41
    adesh
    -1.41
    POSITIVE LOGITS
    cdn
    1.65
    tech
    1.56
     window
    1.55
    blogger
    1.49
    film
    1.47
    jee
    1.46
    GEN
    1.43
    youtube
    1.43
    facebook
    1.41
    world
    1.40
    Activations Density 0.811%

    No Known Activations