INDEX

Explanations

instances of the word "an" followed by another word, potentially indicating a specific context or relationship between the two words

oai_token-act-pair · gpt-3.5-turbo

phrases related to relationships and personal experiences

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

ATURES

-0.71

arning

-0.70

aware

-0.66

bis

-0.64

 Posts

-0.63

amazon

-0.63

yond

-0.62

ensing

-0.62

Former

-0.60

national

-0.60

POSITIVE LOGITS

 conversation

1.08

 conversations

1.07

 meltdown

1.04

 chance

1.02

iphany

1.01

 affair

0.98

 dealings

0.95

fun

0.93

 luck

0.92

 qual

0.90

Activations Density 0.205%

instances of the word "an" followed by another word, potentially indicating a specific context or relationship between the two words

phrases related to relationships and personal experiences

No Comments

No Known Activations