© 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. 12745
    Prev
    Next
    INDEX
    Explanations

    negative or dismissive terms expressing absence or lack

    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
    247
    +0.14
    0.8%
    59
    +0.12
    0.7%
    100
    +0.12
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    239
    +0.14
    0.03
    360
    +0.12
    0.02
    100
    +0.12
    0.02
    Negative Logits
    omitempty
    -1.67
    las
    -1.64
    unden
    -1.52
     millions
    -1.50
    APTER
    -1.48
     lots
    -1.43
    ula
    -1.43
    oral
    -1.40
     alot
    -1.38
    ungen
    -1.37
    POSITIVE LOGITS
    quilibrium
    2.28
    theless
    1.88
    qa
    1.61
     whatsoever
    1.57
     else
    1.53
    pile
    1.47
    where
    1.44
    oft
    1.43
    owing
    1.43
    wegian
    1.40
    Activations Density 0.122%

    No Known Activations