INDEX

Explanations

questions or expressions of curiosity related to location or attribution

oai_token-act-pair · gpt-3.5-turbo

inquiries using the word "where" and variations of the verb "to be."

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

peat

-0.70

UTH

-0.66

legram

-0.66

QUIRE

-0.62

force

-0.61

-0.59

rette

-0.59

factor

-0.58

 Thor

-0.57

atform

-0.57

POSITIVE LOGITS

anwhile

0.64

olation

0.62

 internet

0.60

BuyableInstoreAndOnline

0.60

inois

0.59

ãĤ»

0.58

 annotation

0.58

inventoryQuantity

0.58

DragonMagazine

0.56

adian

0.56

Activations Density 0.046%

questions or expressions of curiosity related to location or attribution

inquiries using the word "where" and variations of the verb "to be."

No Comments

No Known Activations