© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    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. 4-RES-SM
    11. 12003
    Prev
    Next
    INDEX
    Explanations

    expressions and phrases related to personal feelings and relationships

    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
    +0.18
    1.1%
    478
    +0.16
    1.0%
    56
    +0.14
    0.9%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    56
    +0.18
    0.05
    263
    +0.16
    0.12
    498
    +0.14
    0.12
    Negative Logits
    ĥ½
    -2.31
    ↵
    -2.16
    ↵
    -2.15
    -2.15
    -2.15
    ↵
    -2.15
    <|outofrange|>
    -2.15
    -2.15
    -2.15
    -2.15
    POSITIVE LOGITS
     gotta
    2.52
     gonna
    2.45
    hhhh
    2.42
     shit
    2.36
    hh
    2.31
    eeee
    2.26
     crap
    2.26
    ?!
    2.26
     fuckin
    2.21
     fuck
    2.08
    Activations Density 1.531%

    No Known Activations