INDEX

Explanations

phrases related to locations or entities, potentially emphasizing specific names or labels

oai_token-act-pair · gpt-3.5-turbo

the word "Or."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

ãĤ¼ãĤ¦ãĤ¹

-0.79

è¦ļéĨĴ

-0.77

perature

-0.74

pering

-0.69

 guiActiveUnfocused

-0.66

BILITIES

-0.66

Desktop

-0.65

 Doodle

-0.64

Material

-0.62

masters

-0.62

POSITIVE LOGITS

lando

1.24

thodox

1.20

leans

1.05

Else

0.97

phan

0.95

chard

0.94

phans

0.91

acle

0.89

withstanding

0.88

ific

0.85

Activations Density 0.005%

phrases related to locations or entities, potentially emphasizing specific names or labels

the word "Or."

No Comments

No Known Activations