INDEX

Explanations

direct speech quotations

oai_token-act-pair · gpt-3.5-turbo

instances of dialogue or quotation markers in the text, specifically the word "said."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

xtap

-0.74

cv

-0.72

wx

-0.68

folios

-0.67

ritional

-0.65

pleting

-0.65

=~=~

-0.65

clad

-0.64

mental

-0.64

ãĥ´

-0.63

POSITIVE LOGITS

 goodbye

1.43

 hello

1.13

 aloud

0.99

 nothing

0.91

bye

0.91

 farewell

0.87

 Goodbye

0.80

yes

0.78

 bluntly

0.78

hi

0.75

Activations Density 0.183%

direct speech quotations

instances of dialogue or quotation markers in the text, specifically the word "said."

No Comments

No Known Activations