INDEX

Explanations

phrases related to providing insight or information

oai_token-act-pair · gpt-3.5-turbo

expressions of insight or understanding about a subject

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

wa

-0.74

iki

-0.72

soever

-0.72

ICT

-0.70

eah

-0.69

nc

-0.65

ickle

-0.64

glomer

-0.64

itch

-0.63

aughters

-0.62

POSITIVE LOGITS

 what

1.40

how

1.35

 whats

1.33

why

1.31

WHY

1.06

what

0.95

why

0.94

HOW

0.87

 whence

0.86

 sorts

0.84

Activations Density 0.232%

phrases related to providing insight or information

expressions of insight or understanding about a subject

No Comments

No Known Activations