© 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. 10378
    Prev
    Next
    INDEX
    Explanations

    mentions of photographic credits and image sources

    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
    179
    +0.15
    0.8%
    21
    +0.12
    0.7%
    51
    +0.12
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    111
    +0.15
    0.20
    156
    +0.12
    0.16
    71
    +0.12
    0.15
    Negative Logits
    ?",
    -1.67
    ifier
    -1.67
    itud
    -1.64
     heretofore
    -1.57
    essee
    -1.53
     together
    -1.53
     lately
    -1.52
    ifiers
    -1.50
     defined
    -1.49
     satisf
    -1.47
    POSITIVE LOGITS
    UTERS
    3.17
     Getty
    3.06
     AFP
    3.00
    AFP
    2.85
     Reuters
    2.68
     Photo
    2.64
    Getty
    2.62
     Caption
    2.53
    Reuters
    2.48
     photo
    2.45
    Activations Density 1.229%

    No Known Activations