© 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. 14383
    Prev
    Next
    INDEX
    Explanations

    chemical compounds and biochemical processes related to enzymatic functions

    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
    +0.21
    1.2%
    289
    +0.14
    0.8%
    11
    +0.12
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    289
    +0.21
    0.02
    11
    +0.14
    0.03
    448
    +0.12
    0.03
    Negative Logits
    ontally
    -1.74
    surgical
    -1.67
    fits
    -1.65
    ifold
    -1.62
    RECT
    -1.59
    ULAR
    -1.53
    oscope
    -1.47
    orems
    -1.45
    érie
    -1.44
    ediatric
    -1.44
    POSITIVE LOGITS
    «
    2.58
    ¸
    2.45
    °
    2.44
    ¹
    2.33
    §
    2.31
    ·
    2.30
    ĥ
    2.28
    Ĩ
    2.18
     phosphorylation
    2.09
     phosphory
    2.04
    Activations Density 0.085%

    No Known Activations