INDEX

Explanations

phrases related to transitioning from one state or location to another

oai_token-act-pair · gpt-3.5-turbo

phrases indicating transitions or movements between two entities or states

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

lees

-0.65

ãģ®å®

-0.64

ONSORED

-0.60

CLASSIFIED

-0.59

enegger

-0.59

ree

-0.56

ibr

-0.55

urch

-0.54

================================================================

-0.52

 bribes

-0.52

POSITIVE LOGITS

 another

0.90

 Another

0.75

another

0.75

 Tycoon

0.66

 others

0.64

 Confederation

0.63

isine

0.63

rium

0.61

one

0.61

ername

0.61

Activations Density 0.303%

phrases related to transitioning from one state or location to another

phrases indicating transitions or movements between two entities or states

No Comments

No Known Activations