© 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. 14670
    Prev
    Next
    INDEX
    Explanations

    phrases related to financial costs and considerations

    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
    306
    +0.36
    2.6%
    307
    +0.30
    2.2%
    21
    +0.22
    1.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    450
    +0.36
    0.15
    21
    +0.30
    0.14
    306
    +0.22
    0.15
    Negative Logits
    Ļª
    -1.77
    µ
    -1.56
    »¿
    -1.55
     thereof
    -1.54
    lov
    -1.43
    ¿½
    -1.43
    ħ
    -1.41
    Į
    -1.40
    ¨
    -1.40
    Ĩ
    -1.38
    POSITIVE LOGITS
    xhtml
    1.45
    mozilla
    1.40
    className
    1.32
     »
    1.27
    ament
    1.25
    pmed
    1.22
    msgid
    1.22
    ActivityThread
    1.21
    Microsoft
    1.21
    conviction
    1.18
    Activations Density 0.750%

    No Known Activations