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. 1-RES-JB
    6. 15365
    Prev
    Next
    INDEX
    Explanations

    adjectives describing violence or aggression

    oai_token-act-pair · gpt-3.5-turbo

    instances of the word "vicious" and related attributes

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 1-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    1
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    ITNESS
    -0.86
    mberg
    -0.83
    ylon
    -0.83
    avez
    -0.80
    inoa
    -0.77
    agine
    -0.76
    brance
    -0.76
    edia
    -0.75
     Hemp
    -0.72
    irk
    -0.71
    POSITIVE LOGITS
     vicious
    1.06
    ly
    1.03
    nesses
    0.87
     adolesc
    0.81
    iously
    0.80
     savage
    0.80
     earthqu
    0.79
    ness
    0.77
     beasts
    0.76
     thug
    0.74
    Activations Density 0.010%

    No Known Activations