INDEX

Explanations

instances of the word "think" within statements

oai_token-act-pair · gpt-3.5-turbo

expressions of opinion or belief

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

artney

-0.81

ategory

-0.74

clad

-0.72

 clad

-0.71

 flats

-0.67

=~=~

-0.66

 Adin

-0.64

announced

-0.64

 deposited

-0.61

ancers

-0.61

POSITIVE LOGITS

ĸ

0.77

ĳå£«

0.77

eteen

0.73

\\\\\\\\

0.71

ãĤ¦ãĤ¹

0.71

^^^^

0.70

ventus

0.68

itia

0.68

 caucuses

0.68

Activations Density 0.052%

instances of the word "think" within statements

expressions of opinion or belief

No Comments

No Known Activations