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

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Joseph Bloom · Open Source Sparse Autoencoders for all Residual Stream Layers of GPT2-Small
    3. GPT2-Small
    4. Residual Stream
    5. 10-RES-JB
    6. 14564
    Prev
    Next
    INDEX
    Explanations

    phrases with a negative connotation, particularly focusing on the word "no"

    oai_token-act-pair · gpt-3.5-turbo

    negations or phrases indicating absence or denial

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 10-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.hook_resid_pre
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    24,576
    Data Type
    torch.float32
    Hook Point
    blocks.10.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    10
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    rn
    -0.80
    often
    -0.72
    inarily
    -0.71
    ellect
    -0.68
    alian
    -0.67
    ahime
    -0.67
    deck
    -0.66
    schild
    -0.66
     WATCHED
    -0.66
    typically
    -0.65
    POSITIVE LOGITS
     exceptions
    1.11
     longer
    0.99
    xious
    0.97
     modifications
    0.95
     alteration
    0.93
     refunds
    0.91
     compromises
    0.90
     restrictions
    0.89
     surprises
    0.89
     additional
    0.88
    Activations Density 0.106%

    No Known Activations