© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    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. 2-RES-SM
    11. 12547
    Prev
    Next
    INDEX
    Explanations

    adjectives related to positivity and productivity

    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
    232
    +0.12
    0.7%
    362
    +0.12
    0.7%
    187
    +0.11
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    232
    +0.12
    0.04
    445
    +0.12
    0.01
    504
    +0.11
    0.02
    Negative Logits
    chitz
    -1.63
    iqu
    -1.52
    owed
    -1.49
     photographed
    -1.48
    icolor
    -1.47
    otte
    -1.46
    anse
    -1.43
     metro
    -1.41
     numbered
    -1.36
    roe
    -1.35
    POSITIVE LOGITS
    ctors
    1.76
    commons
    1.75
    backs
    1.66
     Commons
    1.63
    minded
    1.61
    ytics
    1.61
    ption
    1.54
    ence
    1.53
    glass
    1.52
     basis
    1.49
    Activations Density 0.241%

    No Known Activations