INDEX

Explanations

the word 'lining' or 'lined'

oai_token-act-pair · gpt-3.5-turbo

instances of the word "line" in various forms

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

 guiIcon

-0.81

arb

-0.78

reck

-0.72

 cumbers

-0.67

è¦ļéĨĴ

-0.66

abuse

-0.65

ILY

-0.65

ppa

-0.65

illac

-0.65

ULT

-0.65

POSITIVE LOGITS

 lining

0.92

lines

0.78

 lined

0.78

ength

0.73

backer

0.73

ups

0.69

 Lines

0.68

breaks

0.68

tsky

0.68

naire

0.68

Activations Density 0.007%

the word 'lining' or 'lined'

instances of the word "line" in various forms

No Comments

No Known Activations