INDEX

Explanations

reasons or explanations indicated in a sentence

oai_token-act-pair · gpt-3.5-turbo

phrases that explain reasons or justifications for various situations

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

 puck

-0.73

 Carbuncle

-0.71

inav

-0.68

 sweats

-0.66

 helicop

-0.65

 broom

-0.64

chron

-0.63

KY

-0.62

 Samurai

-0.61

eg

-0.60

POSITIVE LOGITS

why

1.16

abl

1.01

WHY

0.99

why

0.93

 behind

0.78

forward

0.73

Why

0.72

Why

0.71

 justifying

0.70

¿½

0.70

Activations Density 0.026%

reasons or explanations indicated in a sentence

phrases that explain reasons or justifications for various situations

No Comments

No Known Activations