© 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. 13848
    Prev
    Next
    INDEX
    Explanations

    references to file paths or libraries in a programming context

    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
    326
    +0.20
    1.1%
    161
    +0.15
    0.9%
    26
    +0.14
    0.8%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    26
    +0.20
    0.07
    326
    +0.15
    0.07
    185
    +0.14
    0.06
    Negative Logits
    ¥
    -1.48
     drafts
    -1.48
    Ļª
    -1.43
    aka
    -1.40
    ially
    -1.37
     charts
    -1.34
    £
    -1.34
    ),$$
    -1.31
    lod
    -1.30
    ipart
    -1.30
    POSITIVE LOGITS
     coma
    1.43
     Depression
    1.39
    belt
    1.34
    â
    1.34
    wart
    1.32
     feud
    1.31
    osus
    1.27
    cake
    1.27
     veget
    1.26
    olation
    1.24
    Activations Density 3.212%

    No Known Activations