© 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. 9688
    Prev
    Next
    INDEX
    Explanations

    phrases expressing desire or preference

    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
    48
    +0.13
    0.7%
    52
    +0.13
    0.7%
    478
    +0.13
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    323
    +0.13
    0.03
    48
    +0.13
    0.03
    159
    +0.13
    0.03
    Negative Logits
    ubert
    -1.67
    į
    -1.66
    etics
    -1.56
    esian
    -1.55
    âĢIJ
    -1.50
    ?**
    -1.48
    feas
    -1.47
    ,**
    -1.45
    ural
    -1.43
    **~
    -1.43
    POSITIVE LOGITS
     nothing
    1.63
    enance
    1.58
     stamp
    1.52
    hold
    1.46
     consistency
    1.45
     credibility
    1.44
     premium
    1.41
     instant
    1.41
     them
    1.40
     safer
    1.39
    Activations Density 0.087%

    No Known Activations