© 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. 31340
    Prev
    Next
    INDEX
    Explanations

    mechanical and engineering-related terms, particularly those related to firearms and their components

    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
    17
    +0.14
    0.8%
    354
    +0.13
    0.7%
    406
    +0.11
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    354
    +0.14
    0.10
    401
    +0.13
    0.08
    434
    +0.11
    0.09
    Negative Logits
    ¾
    -3.63
    ļ
    -3.46
    ¿
    -3.45
    Ń
    -3.37
    ³
    -3.34
    Ĵ
    -3.33
    ²
    -3.32
    ↵
    -3.29
    ↵            ↵
    -3.29
    -3.29
    POSITIVE LOGITS
    inates
    2.00
     itself
    1.83
    iNdEx
    1.64
    ename
    1.61
    vert
    1.55
     specified
    1.50
     terminals
    1.49
    inals
    1.49
    ipel
    1.48
    roads
    1.47
    Activations Density 1.641%

    No Known Activations