© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI 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. 11016
    Prev
    Next
    INDEX
    Explanations

    instructions related to technical setup and configuration

    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
    156
    +0.18
    1.1%
    369
    +0.16
    0.9%
    161
    +0.12
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    234
    +0.18
    0.12
    263
    +0.16
    0.13
    369
    +0.12
    0.00
    Negative Logits
    ncia
    -2.38
    UX
    -1.61
    orphisms
    -1.54
     [[
    -1.48
    hes
    -1.45
    edes
    -1.44
     reciproc
    -1.42
    )]{}
    -1.35
    ÑģÑĤво
    -1.33
    eness
    -1.33
    POSITIVE LOGITS
    jee
    1.52
     hoping
    1.43
     expecting
    1.42
     again
    1.39
     office
    1.36
    zol
    1.35
    EDIT
    1.31
    Figure
    1.30
     gardens
    1.28
     concerts
    1.28
    Activations Density 4.168%

    No Known Activations