INDEX

Explanations

phrases indicating physical positions or locations

oai_token-act-pair · gpt-3.5-turbo

phrases related to standing or positioning

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

catentry

-0.74

batch

-0.62

 Pelicans

-0.62

vous

-0.61

NT

-0.61

haus

-0.60

esan

-0.59

<-

-0.58

gas

-0.58

maps

-0.58

POSITIVE LOGITS

 doors

0.72

 pedest

0.71

 desks

0.69

 principled

0.67

ailability

0.67

ailable

0.65

akable

0.65

oin

0.64

ardless

0.63

atoon

0.63

Activations Density 0.093%

phrases indicating physical positions or locations

phrases related to standing or positioning

No Comments

No Known Activations