INDEX

Explanations

phrases related to spoken or written remarks made by individuals

oai_token-act-pair · gpt-3.5-turbo

statements and comments about public figures or controversial topics

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

ccording

-0.70

tis

-0.66

keys

-0.66

bid

-0.65

 Projects

-0.64

VID

-0.63

vu

-0.62

JV

-0.61

locked

-0.59

PG

-0.59

POSITIVE LOGITS

 about

1.18

 uttered

1.01

 dispar

1.00

 implying

0.99

 aloud

0.98

 regarding

0.98

 ABOUT

0.89

 pertaining

0.88

 praising

0.87

 referring

0.87

Activations Density 0.089%

phrases related to spoken or written remarks made by individuals

statements and comments about public figures or controversial topics

No Comments

No Known Activations