© 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. 30864
    Prev
    Next
    INDEX
    Explanations

    references to physical touch and interactions

    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
    263
    +0.22
    1.3%
    23
    +0.12
    0.8%
    161
    +0.12
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    263
    +0.22
    0.15
    6
    +0.12
    0.12
    23
    +0.12
    0.15
    Negative Logits
    ICT
    -1.80
    CMC
    -1.58
     membership
    -1.57
     premises
    -1.52
     central
    -1.51
     fact
    -1.50
    CT
    -1.49
    OPLE
    -1.48
     regional
    -1.47
     ongoing
    -1.45
    POSITIVE LOGITS
    Ĺ
    1.62
     pads
    1.56
    ovo
    1.55
     gloves
    1.53
    chitz
    1.52
    erved
    1.52
     lips
    1.51
    stood
    1.50
    zzles
    1.47
     glove
    1.45
    Activations Density 3.863%

    No Known Activations