INDEX

Explanations

words related to positions of authority or control

oai_token-act-pair · gpt-3.5-turbo

words related to announcements and declarations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Dragonbound

-0.72

 Dahl

-0.72

 Beasts

-0.69

 Rowling

-0.68

 Lemon

-0.67

 Penguin

-0.66

 Solitaire

-0.66

 Lamp

-0.66

 Bots

-0.65

 Elves

-0.65

POSITIVE LOGITS

abor

0.95

ased

0.93

ction

0.92

irm

0.90

agog

0.90

asons

0.90

urring

0.89

ission

0.87

vered

0.87

ocal

0.86

Activations Density 0.158%

words related to positions of authority or control

words related to announcements and declarations

No Comments

No Known Activations