INDEX

Explanations

references to actions or events happening behind metaphorical or physical closed doors

oai_token-act-pair · gpt-3.5-turbo

references to the concept of "doors" 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 @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

TY

-0.80

 Vide

-0.73

TX

-0.72

TING

-0.71

CTV

-0.68

olson

-0.66

ting

-0.64

TOD

-0.63

lihood

-0.63

 Hots

-0.63

POSITIVE LOGITS

 doors

1.26

doors

1.14

bell

1.09

door

1.07

pring

1.03

holes

1.02

 gates

0.98

 Doors

0.94

opening

0.93

mith

0.93

Activations Density 0.020%

references to actions or events happening behind metaphorical or physical closed doors

references to the concept of "doors" in various contexts

No Comments

No Known Activations