INDEX

Explanations

mentions of the term "America" or words related to it

oai_token-act-pair · gpt-3.5-turbo

instances of the substring "amer" in words throughout the text

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

ĵĺ

-0.73

pressed

-0.68

staking

-0.68

ŃĶ

-0.65

 Dietary

-0.59

cooked

-0.59

Desktop

-0.59

 intim

-0.58

 succ

-0.58

 beef

-0.58

POSITIVE LOGITS

icans

1.40

ICAN

1.33

ican

1.33

gdala

1.16

sonian

1.12

ilial

1.08

ica

1.05

icas

1.02

amer

0.94

gency

0.93

Activations Density 0.009%

mentions of the term "America" or words related to it

instances of the substring "amer" in words throughout the text

No Comments

No Known Activations