INDEX

Explanations

instances where something is made explicitly clear or evident

oai_token-act-pair · gpt-3.5-turbo

phrases indicating clarity or transparency in communication

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

izons

-0.72

gins

-0.71

umbn

-0.66

pes

-0.65

rection

-0.63

zbek

-0.63

miah

-0.62

inqu

-0.61

sembly

-0.61

inse

-0.61

POSITIVE LOGITS

why

0.93

why

0.75

how

0.74

 enough

0.73

 that

0.72

WHY

0.70

ered

0.67

to

0.67

cut

0.66

 sailing

0.65

Activations Density 0.045%

instances where something is made explicitly clear or evident

phrases indicating clarity or transparency in communication

No Comments

No Known Activations