INDEX

Explanations

phrases where something is emphasized or pointed out

oai_token-act-pair · gpt-3.5-turbo

instances of the word "highlighted."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

thood

-0.76

onew

-0.71

aah

-0.69

shake

-0.69

cas

-0.69

atern

-0.68

ni

-0.68

eties

-0.68

ework

-0.67

atell

-0.66

POSITIVE LOGITS

TextColor

0.82

 highlights

0.82

umenthal

0.81

 highlighting

0.76

OWS

0.72

shire

0.72

 warnings

0.71

 highlighted

0.71

="#

0.70

 Chavez

0.67

Activations Density 0.020%

phrases where something is emphasized or pointed out

instances of the word "highlighted."

No Comments

No Known Activations