© 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. 29755
    Prev
    Next
    INDEX
    Explanations

    mentions of specific city names or references to urban areas

    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
    156
    +0.16
    1.0%
    81
    +0.15
    0.9%
    45
    +0.10
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    81
    +0.16
    0.03
    180
    +0.15
    0.02
    45
    +0.10
    0.02
    Negative Logits
     appropriate
    -1.71
    deal
    -1.68
    ĩ
    -1.65
     miscar
    -1.59
    amate
    -1.55
    romycin
    -1.54
     such
    -1.51
    sible
    -1.47
    Ĥ
    -1.43
    dle
    -1.42
    POSITIVE LOGITS
     Lights
    2.24
    ways
    1.91
    ÃŃa
    1.90
    fan
    1.83
     Clerk
    1.81
    oyer
    1.78
    liga
    1.76
    ella
    1.73
    agna
    1.70
    oya
    1.69
    Activations Density 0.103%

    No Known Activations