Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Under Peer Review · Sparse Autoencoders for Pythia-70M-Deduped
    3. Pythia-70M-Deduped
    4. Residuals
    5. 1-RES-SM
    6. 26852
    Prev
    Next
    INDEX
    Explanations

    references to fat in various contexts

    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/1-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.1.hook_resid_post
    Hook Layer
    1
    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
    115
    +0.15
    0.9%
    361
    +0.13
    0.8%
    269
    +0.10
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    361
    +0.15
    0.02
    269
    +0.13
    0.02
    270
    +0.10
    0.02
    Negative Logits
     same
    -1.75
    UGH
    -1.61
     whisper
    -1.57
    ÑģÑĮ
    -1.52
    quiry
    -1.51
    TRODUCTION
    -1.47
     our
    -1.47
     hers
    -1.43
     teach
    -1.43
     yours
    -1.41
    POSITIVE LOGITS
    esan
    1.78
    ally
    1.69
    ceral
    1.68
     burner
    1.66
    igue
    1.65
    ality
    1.63
    eners
    1.58
    âĢIJ
    1.56
    ball
    1.55
     during
    1.53
    Activations Density 0.070%

    No Known Activations