INDEX

Explanations

phrases related to safe or secure places

oai_token-act-pair · gpt-3.5-turbo

the term "haven" in various contexts related to safety, refuge, and protection

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

upp

-0.66

 thickness

-0.64

 inference

-0.64

yp

-0.63

ahon

-0.63

heed

-0.63

activation

-0.62

lass

-0.62

otype

-0.61

ickers

-0.60

POSITIVE LOGITS

geon

0.93

't

0.87

 Been

0.86

ajor

0.81

cheon

0.79

gotten

0.78

been

0.76

edin

0.75

izen

0.74

GOT

0.74

Activations Density 0.014%

phrases related to safe or secure places

the term "haven" in various contexts related to safety, refuge, and protection

No Comments

No Known Activations