© 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. 27264
    Prev
    Next
    INDEX
    Explanations

    instances of emotional expressions or sentiments

    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
    203
    +0.19
    1.2%
    56
    +0.19
    1.1%
    2
    +0.14
    0.9%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    203
    +0.19
    0.14
    2
    +0.19
    0.13
    191
    +0.14
    0.12
    Negative Logits
    {},
    -1.78
    __,
    -1.68
     exhibits
    -1.63
    à±į
    -1.51
    ington
    -1.51
    ,
    -1.46
    tees
    -1.46
    ellation
    -1.45
    ificates
    -1.41
    nikov
    -1.36
    POSITIVE LOGITS
    omitempty
    1.67
    especially
    1.60
     yes
    1.59
    don
    1.59
     huh
    1.55
     boss
    1.53
     especially
    1.51
    yet
    1.51
     though
    1.50
     please
    1.48
    Activations Density 1.758%

    No Known Activations