INDEX

Explanations

adjectives related to human behavior or characteristics

oai_token-act-pair · gpt-3.5-turbo

concepts related to being good or ethical in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

 narrowing

-0.68

 veins

-0.68

 fracturing

-0.64

Ö¼

-0.63

 settling

-0.63

 vacancy

-0.62

 easing

-0.61

CPC

-0.59

 Hague

-0.59

çīĪ

-0.58

POSITIVE LOGITS

 oneself

0.96

busters

0.75

 inclusive

0.73

quished

0.71

cool

0.69

operative

0.69

 Yourself

0.68

uned

0.68

 lookout

0.67

 productive

0.67

Activations Density 0.369%

adjectives related to human behavior or characteristics

concepts related to being good or ethical in various contexts

No Comments

No Known Activations