INDEX

Explanations

instances of the word "hole" along with words related to physical openings or gaps

oai_token-act-pair · gpt-3.5-turbo

references to physical "holes" in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

cia

-0.79

against

-0.72

uscript

-0.70

NetMessage

-0.67

 livest

-0.67

Bal

-0.66

voice

-0.65

ority

-0.64

yss

-0.62

lux

-0.62

POSITIVE LOGITS

 hole

1.28

 Hole

1.18

 holes

1.17

hole

1.15

 drilled

1.10

dug

0.99

holes

0.96

izons

0.87

pots

0.80

 canyon

0.80

Activations Density 0.028%

instances of the word "hole" along with words related to physical openings or gaps

references to physical "holes" in various contexts

No Comments

No Known Activations