INDEX

Explanations

percentages

oai_token-act-pair · gpt-3.5-turbo

percentages in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

DragonMagazine

-0.72

Attributes

-0.68

etta

-0.66

pload

-0.66

externalActionCode

-0.65

values

-0.65

chn

-0.65

Rid

-0.64

anwhile

-0.64

IDES

-0.64

POSITIVE LOGITS

imet

1.10

 cent

1.01

lein

0.80

enaries

0.70

imeter

0.69

igr

0.69

aire

0.68

rowth

0.68

olson

0.67

urion

0.65

Activations Density 0.015%

percentages

percentages in various contexts

No Comments

No Known Activations