INDEX

Explanations

words related to openness or lack of restriction

oai_token-act-pair · gpt-3.5-turbo

references to "open-ended" concepts or themes

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

 Tsarnaev

-0.74

 Tycoon

-0.74

 Brach

-0.71

 Mons

-0.69

UD

-0.68

 Rouge

-0.66

Ń·

-0.66

 Lumpur

-0.65

 Khan

-0.65

 Lowry

-0.64

POSITIVE LOGITS

ended

1.24

minded

1.22

source

1.22

eyed

1.13

facing

1.09

bodied

1.07

office

1.02

air

1.02

access

1.01

skinned

1.01

Activations Density 0.026%

words related to openness or lack of restriction

references to "open-ended" concepts or themes

No Comments

No Known Activations