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. 0-RES-JB
    6. 7338
    Prev
    Next
    INDEX
    Explanations

    contractions such as "I'd" and "we'd"

    oai_token-act-pair · gpt-3.5-turbo

    the contraction "I'd," indicating a conditional or subjective statement about desire or intention

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 0-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    0
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     distinguishing
    -0.62
     exclusion
    -0.56
     survival
    -0.56
     informational
    -0.54
    assembly
    -0.54
     ç¥ŀ
    -0.54
     servicing
    -0.52
     nexus
    -0.52
     elite
    -0.52
     heating
    -0.52
    POSITIVE LOGITS
    'd
    3.60
    'll
    1.66
    've
    1.40
     wouldn
    1.27
    're
    1.25
     Wouldn
    1.21
    would
    1.20
     hadn
    1.19
     would
    1.19
    'm
    1.13
    Activations Density 0.013%

    No Known Activations