© 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. 30606
    Prev
    Next
    INDEX
    Explanations

    technical instructions or processes related to software or programming

    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
    12
    +0.17
    1.0%
    465
    +0.14
    0.8%
    427
    +0.12
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    465
    +0.17
    0.11
    12
    +0.14
    0.10
    292
    +0.12
    0.08
    Negative Logits
    º
    -2.45
    Ļª
    -2.40
    ľĵ
    -2.40
    ±
    -2.29
    ¾
    -2.21
    ½
    -2.18
    ĨĴ
    -2.18
    ®
    -2.17
    Ĥ¬
    -2.15
    ª
    -2.11
    POSITIVE LOGITS
     however
    2.21
     reach
    1.78
    ordo
    1.73
    inted
    1.54
     reached
    1.52
     scope
    1.50
    scope
    1.48
    ourse
    1.48
     meanwhile
    1.47
    sight
    1.42
    Activations Density 2.166%

    No Known Activations