INDEX

Explanations

nicknames or terms used to refer to specific individuals

oai_token-act-pair · gpt-3.5-turbo

references to nicknames and their origins

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

actions

-0.73

 WARN

-0.72

Interstitial

-0.71

ebook

-0.69

EMS

-0.68

oval

-0.67

gressive

-0.66

ruct

-0.65

ressive

-0.65

UI

-0.65

POSITIVE LOGITS

 nickname

1.11

 nick

0.87

 nicknamed

0.80

 moniker

0.78

laus

0.76

names

0.75

named

0.73

 haunt

0.71

 bestowed

0.71

 coined

0.71

Activations Density 0.022%

nicknames or terms used to refer to specific individuals

references to nicknames and their origins

No Comments

No Known Activations