INDEX

Explanations

the word "ore" preceded or succeeded by adjacent characters

oai_token-act-pair · gpt-3.5-turbo

words or fragments related to character names or titles

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

ulhu

-0.79

ottesville

-0.78

reau

-0.76

ccording

-0.73

aneers

-0.67

«ĺ

-0.67

ģĸ

-0.64

adium

-0.63

ographed

-0.63

lehem

-0.63

POSITIVE LOGITS

phrine

0.68

ayn

0.67

 Tribunal

0.64

lander

0.62

ilus

0.60

zi

0.60

 Resort

0.59

Viz

0.58

 Rumble

0.58

Menu

0.57

Activations Density 0.358%

the word "ore" preceded or succeeded by adjacent characters

words or fragments related to character names or titles

No Comments

No Known Activations