INDEX

Explanations

personal interactions and emotions, specifically focused on making someone feel better or understanding their feelings

oai_token-act-pair · gpt-3.5-turbo

emotional expressions regarding personal well-being and interpersonal relationships

oai_token-act-pair · gpt-4o-mini

Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Negative Logits

é¾

-0.71

æ©Ł

-0.69

senal

-0.66

USS

-0.64

 misunder

-0.63

Notable

-0.63

vre

-0.60

FactoryReloaded

-0.60

ussia

-0.59

qqa

-0.59

POSITIVE LOGITS

 knowing

1.07

 Because

1.04

 lest

0.99

 realizing

0.98

 because

0.96

Cause

0.94

because

0.93

cause

0.93

 Especially

0.91

Because

0.86

Activations Density 0.337%

personal interactions and emotions, specifically focused on making someone feel better or understanding their feelings

emotional expressions regarding personal well-being and interpersonal relationships

No Known Activations