INDEX

Explanations

the word "null" or related terms, suggesting a focus on null values or invalidating actions

oai_token-act-pair · gpt-3.5-turbo

terms related to nullification or invalidation

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

 exha

-0.82

 millenn

-0.81

Downloadha

-0.79

 livest

-0.77

BOOK

-0.77

 ingred

-0.74

 toget

-0.74

DonaldTrump

-0.71

ilitating

-0.68

ħĭ

-0.67

POSITIVE LOGITS

ifying

1.21

ified

1.20

ptr

1.19

ifies

1.16

ification

1.12

ities

1.05

ify

1.01

ifier

0.97

eenth

0.96

ity

0.95

Activations Density 0.018%

the word "null" or related terms, suggesting a focus on null values or invalidating actions

terms related to nullification or invalidation

No Comments

No Known Activations