© 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. 12608
    Prev
    Next
    INDEX
    Explanations

    statements of assertion or opinion

    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
    478
    +0.19
    1.1%
    115
    +0.12
    0.7%
    2
    +0.11
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    219
    +0.19
    0.04
    63
    +0.12
    0.03
    115
    +0.11
    0.03
    Negative Logits
    į
    -2.80
    ħ
    -2.58
    ¡
    -2.50
    ĭ
    -2.45
    ¬
    -2.40
    ¹
    -2.33
    °
    -2.33
    Ĥ¬
    -2.32
    ŀ
    -2.31
    ¾
    -2.26
    POSITIVE LOGITS
    hift
    2.04
     missing
    1.68
     lack
    1.65
     goodbye
    1.65
     same
    1.56
     detailed
    1.45
     utility
    1.43
     cref
    1.43
     inclusion
    1.42
     fake
    1.39
    Activations Density 0.252%

    No Known Activations