INDEX

Explanations

the word "flat" combined with intensity-related adjectives, pointing towards a focus on flatness

oai_token-act-pair · gpt-3.5-turbo

references to flat objects or surfaces

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 6-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.6.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.6.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

IVERS

-0.95

EMENT

-0.91

CRIP

-0.86

GBT

-0.83

emonic

-0.78

externalActionCode

-0.77

ymes

-0.74

 appropri

-0.73

 ANGEL

-0.73

Æ

-0.72

POSITIVE LOGITS

tered

1.21

bed

1.18

lining

1.13

lined

1.12

bread

1.06

ulence

1.01

ulent

1.01

tering

0.99

sheets

0.94

ric

0.92

Activations Density 0.022%

the word "flat" combined with intensity-related adjectives, pointing towards a focus on flatness

references to flat objects or surfaces

No Comments

No Known Activations