© 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. 31390
    Prev
    Next
    INDEX
    Explanations

    questions that begin with "Are" or "Do"

    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
    367
    +0.15
    0.9%
    435
    +0.14
    0.8%
    225
    +0.12
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    367
    +0.15
    0.03
    214
    +0.14
    0.03
    435
    +0.12
    0.03
    Negative Logits
    §
    -3.56
    ĨĴ
    -3.15
    ¯
    -3.09
    ®
    -3.09
    Ļª
    -3.08
    º
    -3.06
    ģ
    -3.01
    ©
    -2.97
    »
    -2.93
    ¼
    -2.83
    POSITIVE LOGITS
    wolf
    2.17
     those
    1.76
     AUTH
    1.72
    zzo
    1.71
    xtures
    1.65
     bets
    1.54
     Types
    1.52
     yours
    1.50
     constants
    1.50
    tha
    1.48
    Activations Density 0.324%

    No Known Activations