© 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. 27652
    Prev
    Next
    INDEX
    Explanations

    terms and phrases related to data transmission and measuring signals in networks

    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
    395
    +0.12
    0.7%
    28
    +0.10
    0.6%
    379
    +0.10
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    407
    +0.12
    0.14
    33
    +0.10
    0.17
    455
    +0.10
    0.16
    Negative Logits
    ¾
    -1.60
    ness
    -1.59
     centre
    -1.52
     crunch
    -1.51
     center
    -1.47
     snow
    -1.45
     started
    -1.42
     dice
    -1.42
     Kids
    -1.41
     accent
    -1.41
    POSITIVE LOGITS
    rely
    2.28
    ursuant
    1.98
     foregoing
    1.97
    entially
    1.93
     embodiments
    1.92
     herein
    1.90
     heretofore
    1.89
     Examples
    1.85
    extend
    1.81
    requently
    1.77
    Activations Density 4.588%

    No Known Activations