INDEX

Explanations

mentions of the word "clean"

oai_token-act-pair · gpt-3.5-turbo

variations of the word "clear" in different contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

 Krug

-0.72

Tul

-0.71

 Kings

-0.67

 Paramount

-0.67

 Galaxy

-0.67

 Physicians

-0.66

 Ducks

-0.65

BuyableInstoreAndOnline

-0.64

 Hunters

-0.63

PBS

-0.63

POSITIVE LOGITS

arest

1.06

cle

1.06

avage

0.93

combe

0.92

veland

0.85

ogly

0.84

bage

0.81

fruit

0.75

utic

0.73

xp

0.72

Activations Density 0.006%

mentions of the word "clean"

variations of the word "clear" in different contexts

No Comments

No Known Activations