© 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. 4-RES-SM
    11. 14565
    Prev
    Next
    INDEX
    Explanations

    phrases that emphasize the importance of protection and responsibility

    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/4-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.4.hook_resid_post
    Hook Layer
    4
    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
    +0.25
    1.5%
    307
    +0.22
    1.3%
    137
    +0.16
    1.0%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    436
    +0.25
    0.14
    291
    +0.22
    0.11
    137
    +0.16
    0.15
    Negative Logits
    ãģķãĤĮ
    -1.86
    ÑĬÑĢ
    -1.82
    ãĤīãĤĮ
    -1.81
    ãģķãĤĮãģŁ
    -1.72
    ήÏĤ
    -1.72
    ি
    -1.72
    à±į
    -1.66
    à§
    -1.64
    redients
    -1.64
     constructs
    -1.61
    POSITIVE LOGITS
     chance
    2.45
     lot
    2.34
     few
    2.29
     bit
    2.21
     little
    2.20
     thousand
    2.15
     nicer
    2.11
     decent
    2.07
     bigger
    2.00
     nice
    1.98
    Activations Density 1.243%

    No Known Activations