© 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. 1-RES-SM
    11. 25712
    Prev
    Next
    INDEX
    Explanations

    references to Iran and related entities or contexts

    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/1-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.1.hook_resid_post
    Hook Layer
    1
    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
    +0.21
    1.2%
    71
    +0.17
    1.0%
    63
    +0.13
    0.8%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    71
    +0.21
    0.04
    327
    +0.17
    0.03
    279
    +0.13
    0.03
    Negative Logits
    ĥ½
    -1.73
    °
    -1.63
     punch
    -1.60
    uring
    -1.59
    viously
    -1.58
    Ļ
    -1.57
     forth
    -1.53
    ľ
    -1.52
     data
    -1.52
    ancies
    -1.50
    POSITIVE LOGITS
    istan
    2.03
    ians
    2.02
    abad
    2.00
    istani
    1.83
    iÄĩ
    1.78
    nals
    1.72
    nikov
    1.65
    ian
    1.64
     itself
    1.63
     Today
    1.62
    Activations Density 0.189%

    No Known Activations