INDEX

Explanations

verbs related to enjoying an experience or activity

oai_token-act-pair · gpt-3.5-turbo

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

•GPT2 Layer 8 Movie Sentiment Features

No Comments

Negative Logits

ural

-0.73

rawn

-0.64

pled

-0.62

 scrimmage

-0.62

soDeliveryDate

-0.61

ethical

-0.60

tra

-0.58

trace

-0.57

 dare

-0.57

gap

-0.57

POSITIVE LOGITS

ably

0.97

lihood

0.95

 enjoyment

0.81

 watching

0.79

OWER

0.76

 enjoying

0.74

ĸļ

0.73

 chatting

0.71

 Flavoring

0.70

 browsing

0.70

Activations Density 0.053%

verbs related to enjoying an experience or activity

No Comments

No Known Activations