INDEX

Explanations

phrases related to conveying information or messages

oai_token-act-pair · gpt-3.5-turbo

prepositions and words indicating direction or location

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

 Apex

-0.71

yles

-0.71

ãĤ´ãĥ³

-0.68

rower

-0.68

ILY

-0.65

Wyn

-0.65

 è£ıè¦ļéĨĴ

-0.63

achy

-0.62

QL

-0.61

ascript

-0.60

POSITIVE LOGITS

ordinate

0.84

adr

0.74

virt

0.74

effect

0.67

efficiency

0.65

pri

0.64

digit

0.64

uctions

0.63

prison

0.63

chio

0.63

Activations Density 0.374%

phrases related to conveying information or messages

prepositions and words indicating direction or location

No Comments

No Known Activations