INDEX

Explanations

imperatives or directives

oai_token-act-pair · gpt-3.5-turbo

exclamatory statements or punctuation indicating strong emotion or surprise

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

matically

-0.75

oria

-0.74

ching

-0.72

eling

-0.72

der

-0.71

ning

-0.69

med

-0.67

bor

-0.67

washing

-0.67

otto

-0.67

POSITIVE LOGITS

CLASSIFIED

0.89

?!

0.86

#$

0.85

 Anyway

0.83

;)

0.82

 Cohn

0.81

 ðŁĻĤ

0.76

:-)

0.74

!)

0.73

 Marino

0.71

Activations Density 0.016%

imperatives or directives

exclamatory statements or punctuation indicating strong emotion or surprise

No Comments

No Known Activations