INDEX

Explanations

numerical tokens in a specific format

oai_token-act-pair · gpt-3.5-turbo

the end-of-document tokens or markers indicating the conclusion of a text

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 6-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.6.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.6.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

 highs

-0.72

 Cerberus

-0.71

 horizont

-0.71

 breeze

-0.71

Ĥİ

-0.69

 Thumbnails

-0.67

 multiplying

-0.65

 downed

-0.65

 swings

-0.65

wip

-0.64

POSITIVE LOGITS

uggets

1.25

erves

1.14

guyen

1.08

ucle

1.07

umerous

1.07

aughty

1.06

ominated

1.05

omin

1.04

elson

1.04

ihil

1.04

Activations Density 0.031%

numerical tokens in a specific format

the end-of-document tokens or markers indicating the conclusion of a text

No Comments

No Known Activations