INDEX

Explanations

mentions of "all" occurring at very high or high activation levels

oai_token-act-pair · gpt-3.5-turbo

phrases emphasizing the concept of "all."

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

•All the Features!

No Comments

Negative Logits

SHIP

-0.73

heet

-0.65

culosis

-0.65

bal

-0.64

ichen

-0.64

 Flavoring

-0.64

atur

-0.63

hz

-0.63

edin

-0.63

mma

-0.63

POSITIVE LOGITS

uding

1.21

ocating

1.08

owing

1.04

iances

1.01

igators

1.00

usions

1.00

 kinds

0.97

 sorts

0.95

ocated

0.90

usion

0.89

Activations Density 0.057%

mentions of "all" occurring at very high or high activation levels

phrases emphasizing the concept of "all."

No Comments

No Known Activations