INDEX

Explanations

words related to knobs and handles

oai_token-act-pair · gpt-3.5-turbo

mentions of "kn" or "Kn" variations, which likely relates to a specific character or term in the content

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

quo

-0.66

Gan

-0.63

grounds

-0.62

many

-0.61

 actors

-0.61

gent

-0.60

 princip

-0.59

Gum

-0.59

hazard

-0.58

sem

-0.58

POSITIVE LOGITS

uckle

1.59

uckles

1.38

obs

1.26

ighth

1.26

ocking

1.23

ocker

1.19

ows

1.18

itting

1.17

IFE

1.15

itted

1.15

Activations Density 0.035%

words related to knobs and handles

mentions of "kn" or "Kn" variations, which likely relates to a specific character or term in the content

No Comments

No Known Activations