INDEX

Explanations

proper nouns in the format "X.X." where X represents letters or abbreviations

oai_token-act-pair · gpt-3.5-turbo

mentions of the character "J" in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

holders

-0.88

holder

-0.75

dayName

-0.70

ãĤ¤ãĥĪ

-0.65

ciation

-0.63

 Soviets

-0.62

mble

-0.62

TPPStreamerBot

-0.61

 priv

-0.61

 captcha

-0.61

POSITIVE LOGITS

upiter

1.27

ournals

1.25

ealous

1.24

unction

1.12

acket

1.11

ihad

1.05

igsaw

1.04

OY

1.03

ansen

1.03

umper

1.01

Activations Density 0.041%

proper nouns in the format "X.X." where X represents letters or abbreviations

mentions of the character "J" in various contexts

No Comments

No Known Activations