© 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. 14738
    Prev
    Next
    INDEX
    Explanations

    XML declaration and schema-related elements

    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
    +0.20
    1.2%
    307
    +0.16
    1.0%
    348
    +0.15
    0.9%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    180
    +0.20
    0.01
    160
    +0.16
    0.01
    46
    +0.15
    0.01
    Negative Logits
    ways
    -1.70
    quarters
    -1.59
     specificity
    -1.47
    hops
    -1.47
     offence
    -1.44
     needed
    -1.41
     hypotheses
    -1.39
    cule
    -1.38
    folk
    -1.37
    rers
    -1.37
    POSITIVE LOGITS
    «
    2.56
    ¦
    2.54
    Ļª
    2.43
    ↵↵
    2.41
    ↵↵↵
    2.41
    ↵↵
    2.41
    2.41
    ↵
    2.41
    ↵
    2.41
    č↵
    2.41
    Activations Density 0.040%

    No Known Activations