© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it
    Neuronpedia
    Jacobian LensNEW
    Natural Language
    Autoencoders
    MAY 26    Assistant AxisJAN 26Circuit TracerAPR 26SteerSAE EvalsExportsinterp-engineNEW 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. 26066
    Prev
    Next
    INDEX
    Explanations

    references to hormones and peptides related to aggression and emotional regulation

    oai_token-act-pair · gpt-4o-mini
    Triggered 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
    Neuron Alignment
    Index
    Value
    % of L₁
    47
    +0.17
    0.9%
    320
    +0.12
    0.7%
    289
    +0.12
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    271
    +0.17
    0.01
    66
    +0.12
    0.04
    47
    +0.12
    0.04
    Negative Logits
     misunderstanding
    -1.78
     vista
    -1.75
     hurry
    -1.59
     Astronomy
    -1.57
     Photoshop
    -1.56
     frustration
    -1.53
     Virgin
    -1.53
     Places
    -1.50
     ÑĩÑĤо
    -1.50
     misunder
    -1.50
    POSITIVE LOGITS
     administered
    1.90
    forcement
    1.88
     oxidase
    1.84
    oxacin
    1.83
    umab
    1.76
    osis
    1.75
     produced
    1.73
    eral
    1.73
    acillus
    1.72
    oxin
    1.70
    Activations Density 0.232%

    No Known Activations