INDEX

Explanations

statements or opinions expressed in written form

oai_token-act-pair · gpt-3.5-turbo

phrases related to communication, dialogue, and the sharing of information

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

ayne

-0.70

asus

-0.68

capacity

-0.62

audi

-0.62

berus

-0.61

ossus

-0.59

senal

-0.58

 bidder

-0.57

 Roses

-0.57

orest

-0.56

POSITIVE LOGITS

 about

1.54

 ABOUT

1.46

about

1.17

 About

1.14

 aloud

1.14

 anonymously

1.09

 regarding

1.06

 concerning

1.02

 pertaining

0.98

 orally

0.98

Activations Density 0.563%

statements or opinions expressed in written form

phrases related to communication, dialogue, and the sharing of information

No Comments

No Known Activations