© 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. 4-RES-SM
    11. 9393
    Prev
    Next
    INDEX
    Explanations

    product descriptions focused on quality and features of household items or accessories

    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/4-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.4.hook_resid_post
    Hook Layer
    4
    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
    307
    +0.20
    1.1%
    202
    +0.11
    0.6%
    309
    +0.11
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    67
    +0.20
    0.02
    60
    +0.11
    0.05
    490
    +0.11
    0.07
    Negative Logits
     (§
    -2.00
     meetings
    -1.93
    /~
    -1.85
     surveillance
    -1.82
    à±į
    -1.81
    overnment
    -1.78
     conversations
    -1.68
     conferences
    -1.67
     0000000000000000000000000000000000
    -1.67
     IRC
    -1.64
    POSITIVE LOGITS
     stylish
    2.47
     stainless
    2.38
     beautifully
    2.38
     crafted
    2.30
     leather
    2.23
     decorative
    2.21
     customized
    2.20
     Features
    2.16
     customize
    2.15
     versatile
    2.13
    Activations Density 3.711%

    No Known Activations