© 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. 12656
    Prev
    Next
    INDEX
    Explanations

    terms related to ionic substances and their properties

    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
    156
    +0.17
    1.0%
    13
    +0.13
    0.8%
    47
    +0.13
    0.8%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    13
    +0.17
    0.01
    177
    +0.13
    0.01
    275
    +0.13
    0.01
    Negative Logits
     particular
    -1.80
    orkshire
    -1.63
    abouts
    -1.58
     '../
    -1.58
    aurant
    -1.58
     noct
    -1.55
     minds
    -1.52
     mind
    -1.52
     suf
    -1.51
     corner
    -1.50
    POSITIVE LOGITS
    Ģ
    2.26
    naire
    1.78
    ģ
    1.74
     beams
    1.73
    ¤
    1.66
    berg
    1.64
    ī
    1.63
    stadt
    1.62
    pow
    1.59
    rate
    1.58
    Activations Density 0.079%

    No Known Activations