INDEX

Explanations

titles or phrases within quotes

oai_token-act-pair · gpt-3.5-turbo

quotation marks and punctuation related to titles of works or phrases

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

²¾

-0.74

cha

-0.73

¬¼

-0.71

senal

-0.70

ĻĤ

-0.69

nex

-0.67

worldly

-0.67

itionally

-0.67

©¶æ

-0.66

 favourites

-0.64

POSITIVE LOGITS

/"

0.85

 according

0.78

 referring

0.77

>>\

0.75

 meaning

0.73

aka

0.73

 implying

0.73

etc

0.72

SPONSORED

0.71

 namely

0.67

Activations Density 0.070%

titles or phrases within quotes

quotation marks and punctuation related to titles of works or phrases

No Comments

No Known Activations