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

    Neuronpedia

    APIAssistant AxisNEWCircuit 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. 4-RES-JB
    6. 21747
    Prev
    Next
    INDEX
    Explanations

    commands or instructions that include the word "should"

    oai_token-act-pair · gpt-3.5-turbo

    imperative statements or recommendations

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

    No Comments

    Negative Logits
     Puzzle
    -0.65
     Memories
    -0.64
    reality
    -0.64
     Fra
    -0.63
    atile
    -0.63
     Kore
    -0.61
     Syndrome
    -0.61
    cule
    -0.60
    ils
    -0.59
     Blaz
    -0.59
    POSITIVE LOGITS
    ered
    0.97
     be
    0.95
     ideally
    0.95
     clarify
    0.86
    ering
    0.85
     capitalize
    0.83
     definitely
    0.82
     probably
    0.81
     preferably
    0.80
    edi
    0.78
    Activations Density 0.056%

    No Known Activations