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

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsBlogSlackPrivacy & 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. 2-RES-JB
    6. 32
    Prev
    Next
    INDEX
    Explanations

    mentions of the word "melon" or phonetically similar variations like "mel" or "Mell"

    oai_token-act-pair · gpt-3.5-turbo

    mentions of "mel" and its variations, indicating a focus on music and artistic elements related to melody

    oai_token-act-pair · gpt-4o-miniTriggered by @bot

    the syllable 'mel'

    oai_token-act-pair · gpt-4-turbo
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 2-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    2
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     Citizen
    -0.82
     Fit
    -0.74
    GV
    -0.72
    BLIC
    -0.71
     Oath
    -0.71
    æĸ¹
    -0.71
     Fear
    -0.70
     Directors
    -0.70
    ij士
    -0.68
     Citizens
    -0.68
    POSITIVE LOGITS
    atonin
    1.33
    anz
    1.05
    odic
    1.03
    iband
    0.97
    bourg
    0.97
    istors
    0.95
    ient
    0.93
    bourne
    0.92
    mel
    0.91
    astic
    0.91
    Activations Density 0.005%

    No Known Activations