© 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. 4-RES-SM
    11. 9009
    Prev
    Next
    INDEX
    Explanations

    legal terms related to warranties and conditions

    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
    78
    +0.26
    1.7%
    271
    +0.12
    0.8%
    415
    +0.11
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    415
    +0.26
    0.02
    469
    +0.12
    0.02
    165
    +0.11
    0.02
    Negative Logits
    ital
    -1.55
     saf
    -1.34
     engagement
    -1.32
    aries
    -1.31
    mente
    -1.30
    azine
    -1.19
     caliber
    -1.19
     themselves
    -1.19
    enstein
    -1.18
     heads
    -1.18
    POSITIVE LOGITS
    Ī
    3.88
    Ħ
    3.87
    ŀ
    3.81
    ģ
    3.78
    ¸
    3.76
    Īĺ
    3.70
    ī
    3.65
    Ļª
    3.55
    į
    3.53
    ¼
    3.52
    Activations Density 0.061%

    No Known Activations