© 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. 13702
    Prev
    Next
    INDEX
    Explanations

    phrases or expressions indicating a significant quantity or abundance

    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
    222
    +0.11
    0.6%
    127
    +0.10
    0.6%
    102
    +0.10
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    375
    +0.11
    0.04
    222
    +0.10
    0.04
    102
    +0.10
    0.04
    Negative Logits
    (_
    -1.79
    ¿
    -1.67
    ETHOD
    -1.63
    documentclass
    -1.59
    aphragm
    -1.58
    curs
    -1.57
    disambiguation
    -1.53
    ľĵ
    -1.51
    Ĩ
    -1.48
     assassination
    -1.45
    POSITIVE LOGITS
    ergy
    1.60
    quilibrium
    1.52
     wel
    1.46
    igue
    1.44
    inki
    1.39
     wood
    1.38
    heit
    1.37
     oxidized
    1.35
     emphasis
    1.34
    ashi
    1.33
    Activations Density 0.128%

    No Known Activations