INDEX

Explanations

descriptions related to visual appearance

oai_token-act-pair · gpt-3.5-turbo

phrases that contain the word "looks."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

die

-0.74

âĹ¼

-0.70

sov

-0.68

oled

-0.68

death

-0.68

ricular

-0.66

SPONSORED

-0.65

une

-0.65

learning

-0.63

sec

-0.62

POSITIVE LOGITS

peed

0.82

rul

0.80

ynthesis

0.79

 suspic

0.78

":"/

0.76

afety

0.75

ahead

0.74

metics

0.74

earch

0.73

ometimes

0.72

Activations Density 0.025%

descriptions related to visual appearance

phrases that contain the word "looks."

No Comments

No Known Activations