INDEX

Explanations

phrases that indicate an action or decision made by a spokesperson or authority figure

oai_token-act-pair · gpt-3.5-turbo

the word "that" in various contexts

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

ãĤ©

-0.80

ãĤ¼ãĤ¦ãĤ¹

-0.73

greg

-0.69

tein

-0.66

EMBER

-0.66

INFO

-0.66

Tank

-0.65

Ü

-0.65

ãĥĺ

-0.65

Thumbnail

-0.65

POSITIVE LOGITS

 although

1.37

 while

1.20

 despite

1.18

"[

1.15

 whilst

1.04

 whereas

1.02

 unlike

1.02

 unless

1.00

 "âĢ¦

0.92

if

0.90

Activations Density 0.231%

phrases that indicate an action or decision made by a spokesperson or authority figure

the word "that" in various contexts

No Comments

No Known Activations