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. 4-RES-JB
    6. 21435
    Prev
    Next
    INDEX
    Explanations

    words related to severe or intense conditions or situations

    oai_token-act-pair · gpt-3.5-turbo

    instances of the word "harsh" and its associated phrases describing severity or brutality

    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
    ovember
    -1.00
    phis
    -0.98
    ĸļ
    -0.93
    ITNESS
    -0.84
    ablish
    -0.81
    assies
    -0.80
    cellent
    -0.79
    ICLE
    -0.75
    igate
    -0.74
    aintain
    -0.73
    POSITIVE LOGITS
     harsh
    1.15
     harshly
    1.01
     punishments
    0.92
     harsher
    0.91
     hars
    0.90
    ness
    0.90
     punishment
    0.89
    ened
    0.86
     punitive
    0.86
     winters
    0.85
    Activations Density 0.017%

    No Known Activations