© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    1. Home
    3. Joseph Bloom · Open Source Sparse Autoencoders for all Residual Stream Layers of GPT2-Small
    5. GPT2-Small
    7. Residual Stream
    9. 0-RES-JB
    11. 20366
    Prev
    Next
    INDEX
    Explanations

    phrases related to identification or verification, particularly in a professional or security context

    oai_token-act-pair · gpt-3.5-turbo

    references to credentials and related terms

    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
    Decre
    -0.78
    harm
    -0.71
    ranged
    -0.71
     subdiv
    -0.71
     Decre
    -0.69
    Tree
    -0.64
    rez
    -0.64
     Rav
    -0.63
     subdivision
    -0.63
    fur
    -0.63
    POSITIVE LOGITS
     credentials
    3.89
     credential
    2.22
     qualifications
    1.52
     certificates
    1.40
     passwords
    1.33
     privileges
    1.29
     pedigree
    1.27
    entials
    1.26
     bona
    1.24
     certs
    1.23
    Activations Density 0.009%

    No Known Activations