INDEX

Explanations

phrases related to standing or being in a particular position

oai_token-act-pair · gpt-3.5-turbo

phrases related to physical positioning or presence in a location

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

catentry

-0.79

batch

-0.68

expensive

-0.67

processing

-0.64

jing

-0.64

brain

-0.61

NT

-0.60

 headache

-0.60

<-

-0.60

eez

-0.60

POSITIVE LOGITS

 desks

0.74

 STAND

0.69

akable

0.69

oppable

0.68

 banner

0.68

Blocks

0.67

Accessory

0.65

 principled

0.65

ailable

0.65

 doors

0.64

Activations Density 0.095%

phrases related to standing or being in a particular position

phrases related to physical positioning or presence in a location

No Comments

No Known Activations