INDEX

Explanations

words or phrases that underline a point

oai_token-act-pair · gpt-3.5-turbo

the word "obviously."

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

tein

-0.73

arthed

-0.73

lets

-0.71

lake

-0.69

lÃ©

-0.68

nan

-0.67

bucks

-0.67

ictions

-0.66

mission

-0.65

hi

-0.64

POSITIVE LOGITS

icably

0.73

 identifiable

0.71

Ø³

0.71

 misunderstand

0.70

\\\\\\\\

0.70

 Leilan

0.70

NULL

0.70

 correct

0.68

 distinguish

0.68

 audible

0.65

Activations Density 0.011%

words or phrases that underline a point

the word "obviously."

No Comments

No Known Activations