INDEX

Explanations

the word "Boost" with a strong activation

oai_token-act-pair · gpt-3.5-turbo

references to specific software libraries or frameworks

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

Þ

-0.80

ĸļå£«

-0.75

 cryst

-0.73

 tremend

-0.72

 citiz

-0.72

AE

-0.71

 exting

-0.70

ÄŁ

-0.67

pher

-0.66

oded

-0.65

POSITIVE LOGITS

ers

0.92

nuts

0.90

ancers

0.83

ership

0.82

azel

0.81

 Dull

0.80

jamin

0.79

wagen

0.79

 Gleaming

0.78

 Riders

0.77

Activations Density 0.032%

the word "Boost" with a strong activation

references to specific software libraries or frameworks

No Comments

No Known Activations