INDEX

Explanations

words related to categorization or classification, particularly terms related to brackets

oai_token-act-pair · gpt-3.5-turbo

instances of the term "bracket" and its variations

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

natureconservancy

-0.65

 Liberties

-0.64

 Pharmaceutical

-0.63

BuyableInstoreAndOnline

-0.63

Soy

-0.62

 Pharm

-0.61

 Federation

-0.61

Oxy

-0.61

EE

-0.60

 Powerful

-0.60

POSITIVE LOGITS

 brackets

1.57

 bracket

1.50

ackets

1.27

acket

0.94

alon

0.77

halla

0.77

ashtra

0.76

 notation

0.70

prep

0.69

ftime

0.69

Activations Density 0.011%

words related to categorization or classification, particularly terms related to brackets

instances of the term "bracket" and its variations

No Comments

No Known Activations