© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Natural Language
    Autoencoders
    NEW
    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    2. Under Peer Review · Sparse Autoencoders for Pythia-70M-Deduped
    3. Pythia-70M-Deduped
    4. Residuals
    5. 4-RES-SM
    6. 11875
    Prev
    Next
    INDEX
    Explanations

    function calls related to setting configurations or properties

    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
    307
    +0.17
    0.9%
    47
    +0.13
    0.7%
    194
    +0.13
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    17
    +0.17
    0.03
    47
    +0.13
    0.03
    194
    +0.13
    0.02
    Negative Logits
    inus
    -1.51
    Č
    -1.48
    izumab
    -1.47
    UTERS
    -1.41
    gaard
    -1.38
    ickers
    -1.38
    films
    -1.35
    LICENSE
    -1.35
    stra
    -1.34
    INGTON
    -1.34
    POSITIVE LOGITS
    Ħ
    2.94
    upt
    2.69
     up
    2.64
    ı
    2.53
     forth
    2.42
    Timeout
    2.32
    ĥ½
    2.30
    ·¸
    2.28
    ±
    2.24
    ħ
    2.18
    Activations Density 0.165%

    No Known Activations