INDEX

Explanations

words with accents, particularly the letter "é."

oai_token-act-pair · gpt-3.5-turbo

occurrences of the character "é" in various contexts

oai_token-act-pair · gpt-4o-mini Triggered 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

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

utterstock

-0.75

pread

-0.74

enged

-0.73

 Lovecraft

-0.69

berra

-0.68

 Raiders

-0.67

Spread

-0.66

ageddon

-0.66

uyomi

-0.66

 redistributed

-0.65

POSITIVE LOGITS

lÃ©

1.15

 Ã©

0.91

 Dame

0.90

Ã©

0.88

cÃ©

0.87

rique

0.87

vez

0.87

Ã¯

0.86

itÃ©

0.85

rez

0.85

Activations Density 0.011%

words with accents, particularly the letter "é."

occurrences of the character "é" in various contexts

No Comments

No Known Activations