INDEX

Explanations

mentions of stickers, often with high activation values

oai_token-act-pair · gpt-3.5-turbo

references to stickers

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

ernandez

-0.79

MQ

-0.74

ahime

-0.72

acan

-0.72

vernment

-0.70

scill

-0.69

ught

-0.69

abama

-0.68

ITNESS

-0.68

apor

-0.67

POSITIVE LOGITS

 stickers

1.37

 sticker

1.22

holders

0.97

 sheet

0.94

holder

0.84

 stripe

0.83

raft

0.83

sheet

0.82

 markings

0.81

boards

0.78

Activations Density 0.019%

mentions of stickers, often with high activation values

references to stickers

No Comments

No Known Activations