INDEX

Explanations

questions represented by a question mark at the end

oai_token-act-pair · gpt-3.5-turbo

question marks and exclamation points, signaling inquiries or expressions of excitement

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

waukee

-0.71

heny

-0.69

shaw

-0.67

tyard

-0.65

zona

-0.65

esville

-0.65

 imposition

-0.65

inki

-0.64

kefeller

-0.64

ascus

-0.62

POSITIVE LOGITS

ĸļ

0.76

ICLE

0.74

Ħ¢

0.74

 Laure

0.70

0.69

desktop

0.69

FG

0.69

 Cloak

0.67

::::

0.66

POL

0.66

Activations Density 0.008%

questions represented by a question mark at the end

question marks and exclamation points, signaling inquiries or expressions of excitement

No Comments

No Known Activations