INDEX

Explanations

shocking or surprising events or information

oai_token-act-pair · gpt-3.5-turbo

instances of the word "shocking."

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

xia

-0.85

©¶æ

-0.82

 Kinnikuman

-0.80

haps

-0.78

amins

-0.75

ebus

-0.74

ophers

-0.73

bid

-0.71

gran

-0.71

thus

-0.71

POSITIVE LOGITS

 Revelations

0.89

aneously

0.85

 srfAttach

0.80

 shocking

0.80

 revelations

0.78

IFT

0.74

ãĤ¦ãĤ¹

0.74

à¤¾

0.72

 unbeliev

0.72

ãĥŀ

0.71

Activations Density 0.008%

shocking or surprising events or information

instances of the word "shocking."

No Comments

No Known Activations