INDEX

Explanations

items or objects related to physical attributes or characteristics

oai_token-act-pair · gpt-3.5-turbo

descriptors of visual attributes and specific features in objects

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

orthy

-0.87

MQ

-0.87

Financial

-0.83

ocity

-0.82

agonist

-0.80

conservancy

-0.80

obia

-0.80

iphany

-0.80

dule

-0.79

absolute

-0.77

POSITIVE LOGITS

 markings

1.25

 stitching

1.18

 buttons

1.16

 inscription

1.15

 logo

1.15

 accents

1.15

 vents

1.14

 protr

1.14

 triangular

1.12

 zipper

1.12

Activations Density 0.317%

items or objects related to physical attributes or characteristics

descriptors of visual attributes and specific features in objects

No Comments

No Known Activations