INDEX

Explanations

instances where someone expresses interest in something

oai_token-act-pair · gpt-3.5-turbo

expressions of interest or curiosity

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

ework

-0.70

 stacked

-0.68

 fabricated

-0.65

apple

-0.65

car

-0.64

mit

-0.63

 dotted

-0.62

graded

-0.61

 bunk

-0.61

Bom

-0.61

POSITIVE LOGITS

atile

0.75

 Interest

0.75

="#

0.73

 warr

0.72

=#

0.72

 interested

0.72

edIn

0.71

CONT

0.68

enza

0.68

reprene

0.67

Activations Density 0.015%

instances where someone expresses interest in something

expressions of interest or curiosity

No Comments

No Known Activations