INDEX

Explanations

words related to information technology and computing

oai_token-act-pair · gpt-3.5-turbo

terms related to warnings, resources, and systemic issues

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 elves

-0.83

 totem

-0.82

 crest

-0.79

 sciences

-0.78

 contempt

-0.78

 rook

-0.78

 embargo

-0.77

 parity

-0.77

 disg

-0.77

 disarm

-0.77

POSITIVE LOGITS

 Anyway

1.34

 Away

1.34

 Again

1.33

 Regarding

1.31

 Through

1.30

 Than

1.30

 Ahead

1.28

 During

1.28

 About

1.27

 Them

1.27

Activations Density 0.266%

words related to information technology and computing

terms related to warnings, resources, and systemic issues

No Comments

No Known Activations