INDEX

Explanations

words related to itching or irritation

oai_token-act-pair · gpt-3.5-turbo

variations of the word "itch."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

 adjud

-0.66

 custod

-0.66

 lunch

-0.65

deb

-0.65

 magn

-0.64

 Liber

-0.62

 rays

-0.61

 CENT

-0.61

 indirect

-0.60

Kub

-0.60

POSITIVE LOGITS

itching

3.98

itches

3.57

itched

3.23

ITCH

2.66

itch

2.43

itcher

2.28

icked

1.15

inkle

1.07

 hitch

1.04

 Pitch

1.02

Activations Density 0.014%

words related to itching or irritation

variations of the word "itch."

No Comments

No Known Activations