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

    Neuronpedia

    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI 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. 7-RES-JB
    11. 196
    Prev
    Next
    INDEX
    Explanations

    adjectives with negative connotations

    oai_token-act-pair · gpt-3.5-turbo

    descriptive adjectives and phrases that convey strong emotional or provocative qualities

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 7-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    7
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    swick
    -0.74
    essage
    -0.72
    umbn
    -0.69
    anamo
    -0.69
    ailable
    -0.68
    æ©
    -0.61
    avorite
    -0.61
    aldi
    -0.60
    artney
    -0.60
    onds
    -0.60
    POSITIVE LOGITS
     abortions
    0.60
     thinker
    0.57
    \":
    0.56
     eater
    0.54
     thinkers
    0.54
     pronoun
    0.53
     subreddits
    0.52
    poses
    0.51
    adoes
    0.51
     feminist
    0.51
    Activations Density 0.354%

    No Known Activations