© 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 EvalsExportsAPI 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. 30412
    Prev
    Next
    INDEX
    Explanations

    indicators related to emotional struggles and support mechanisms

    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
    293
    +0.20
    1.2%
    12
    +0.17
    1.0%
    263
    +0.15
    0.9%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    293
    +0.20
    0.13
    12
    +0.17
    0.12
    263
    +0.15
    0.13
    Negative Logits
    TypeDef
    -1.69
    hei
    -1.46
    ém
    -1.42
    ubicin
    -1.41
    apon
    -1.35
    cul
    -1.32
     "!"
    -1.30
     "")
    -1.30
     Manual
    -1.29
    ophagus
    -1.28
    POSITIVE LOGITS
    ¤
    1.51
    razil
    1.43
    appellant
    1.37
    ilingual
    1.36
    ]{
    1.34
    ĸ
    1.33
     myself
    1.33
     reviewers
    1.30
    ook
    1.30
    ¼
    1.29
    Activations Density 3.355%

    No Known Activations