INDEX

Explanations

phrases related to high expertise, importance, or intensity

oai_token-act-pair · gpt-3.5-turbo

phrases related to high levels of intensity or evaluation

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

Stars

-0.91

 Stars

-0.86

 Milky

-0.77

 Andromeda

-0.75

 Predators

-0.71

 Tycoon

-0.71

 Steele

-0.69

Pitt

-0.67

 Manny

-0.67

Beast

-0.66

POSITIVE LOGITS

holder

0.80

holders

0.77

 bureaucr

0.74

level

0.73

xual

0.73

lly

0.68

 delegation

0.68

 levels

0.67

 diplomat

0.67

abulary

0.67

Activations Density 0.033%

phrases related to high expertise, importance, or intensity

phrases related to high levels of intensity or evaluation

No Comments

No Known Activations