© 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. 12856
    Prev
    Next
    INDEX
    Explanations

    components related to crafting or artistic creation

    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
    450
    +0.15
    0.9%
    275
    +0.13
    0.7%
    448
    +0.13
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    450
    +0.15
    0.10
    448
    +0.13
    0.04
    140
    +0.13
    0.07
    Negative Logits
    оÐ
    -1.80
    ingle
    -1.77
    Enumerable
    -1.77
    ypse
    -1.75
    frastructure
    -1.72
    OUNT
    -1.68
    PtrFromString
    -1.64
    à°¿
    -1.59
    CDATA
    -1.55
    "}](#
    -1.54
    POSITIVE LOGITS
    ishly
    1.64
    headed
    1.63
     foods
    1.61
    liness
    1.61
    -
    1.60
    zing
    1.56
    ily
    1.54
     glasses
    1.48
    hard
    1.43
    âĢIJ
    1.39
    Activations Density 2.327%

    No Known Activations