© 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 EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Under Peer Review · Sparse Autoencoders for Pythia-70M-Deduped
    5. Pythia-70M-Deduped
    7. Residuals
    9. 1-RES-SM
    11. 25994
    Prev
    Next
    INDEX
    Explanations

    instances of the word "pretty" used to describe something positively

    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/1-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.1.hook_resid_post
    Hook Layer
    1
    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
    255
    +0.14
    0.8%
    444
    +0.12
    0.7%
    376
    +0.12
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    255
    +0.14
    0.04
    272
    +0.12
    0.03
    36
    +0.12
    0.03
    Negative Logits
     times
    -1.49
     permit
    -1.46
    ctomy
    -1.43
    ise
    -1.43
     instance
    -1.41
    ample
    -1.38
    ater
    -1.37
    ^\*^
    -1.37
     time
    -1.35
    enne
    -1.34
    POSITIVE LOGITS
    áĢº
    2.09
    áĢ¬
    1.96
    áĢ¯
    1.95
    ¿½
    1.93
    à±ģ
    1.71
    doing
    1.67
    áĢ¸
    1.67
    prints
    1.63
    stuff
    1.61
    bour
    1.60
    Activations Density 0.095%

    No Known Activations