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. 0-RES-JB
    6. 24404
    Prev
    Next
    INDEX
    Explanations

    the word "deny" in various contexts

    oai_token-act-pair · gpt-3.5-turbo

    instances of the word "deny" and its forms

    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
    cycl
    -0.67
     Caption
    -0.65
    lance
    -0.64
     Instruction
    -0.63
     Toys
    -0.63
     Peg
    -0.62
     Cous
    -0.62
    Poké
    -0.61
     TI
    -0.60
     Ches
    -0.59
    POSITIVE LOGITS
     deny
    3.72
     denies
    2.88
     denying
    2.66
     denied
    2.55
     denial
    2.07
     refuse
    1.36
     deprive
    1.33
     refute
    1.32
     DEN
    1.32
     accuse
    1.31
    Activations Density 0.014%

    No Known Activations