INDEX

Explanations

contractions of "is" or "has" followed by a negative word

oai_token-act-pair · gpt-3.5-turbo

phrases indicating contrasts or negations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

sense

-0.71

cial

-0.68

luaj

-0.66

omever

-0.61

arius

-0.61

idth

-0.60

miah

-0.59

anka

-0.59

ukong

-0.58

ahs

-0.57

POSITIVE LOGITS

©¶æ

0.82

 unlikely

0.81

 irrelevant

0.81

 nowhere

0.79

 offset

0.75

 hardly

0.72

 moot

0.72

still

0.72

 outwe

0.70

 overshadowed

0.70

Activations Density 0.156%

contractions of "is" or "has" followed by a negative word

phrases indicating contrasts or negations

No Comments

No Known Activations