INDEX

Explanations

words related to criticism and controversy

oai_token-act-pair · gpt-3.5-turbo

negative terms and concepts related to conflict and discontent

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

DonaldTrump

-0.79

 Paddock

-0.63

DragonMagazine

-0.61

 Pyramid

-0.56

 Warriors

-0.56

 Kaiser

-0.55

 Trader

-0.55

 House

-0.55

 retailer

-0.55

 "$:/

-0.55

POSITIVE LOGITS

 alike

0.86

 versa

0.78

ject

0.74

interested

0.72

amb

0.71

ilit

0.71

itud

0.70

icing

0.69

ise

0.68

igmat

0.68

Activations Density 0.247%

words related to criticism and controversy

negative terms and concepts related to conflict and discontent

No Comments

No Known Activations