INDEX

Explanations

phrases indicating strong opinions or beliefs

oai_token-act-pair · gpt-3.5-turbo

connections to significant moral or ethical themes

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

 Awakens

-0.56

 Realms

-0.55

 Masquerade

-0.53

Mak

-0.53

 Kenn

-0.52

Sn

-0.52

 Kardash

-0.51

Nik

-0.50

zo

-0.50

 flyers

-0.50

POSITIVE LOGITS

maxwell

0.65

lie

0.62

onite

0.58

ieu

0.57

etheless

0.57

JECT

0.56

Topic

0.56

hement

0.56

oldown

0.56

ŃĶ

0.56

Activations Density 1.440%

phrases indicating strong opinions or beliefs

connections to significant moral or ethical themes

No Comments

No Known Activations