INDEX

Explanations

mentions of the word "attacks" with a high activation value

oai_token-act-pair · gpt-3.5-turbo

mentions of "cats."

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

SOURCE

-0.77

 pall

-0.75

SPONSORED

-0.66

 Quarterly

-0.65

 divest

-0.64

 commencement

-0.62

 curv

-0.62

 planetary

-0.62

©¶æ

-0.61

 concurrent

-0.59

POSITIVE LOGITS

ats

1.35

wana

1.07

terness

1.03

icket

1.03

abase

1.02

herer

0.96

heet

0.96

acus

0.93

htaking

0.92

chers

0.92

Activations Density 0.012%

mentions of the word "attacks" with a high activation value

mentions of "cats."

No Comments

No Known Activations