INDEX

Explanations

references to the location or person named "Win"

oai_token-act-pair · gpt-3.5-turbo

instances of the word "Win" related to various contexts or topics

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

âĶģ

-0.71

¿½

-0.68

ADRA

-0.67

alam

-0.63

 incorpor

-0.63

umn

-0.63

 condu

-0.63

erous

-0.63

gdala

-0.62

 pree

-0.60

POSITIVE LOGITS

ners

1.22

nings

1.20

frey

1.07

throp

1.07

now

0.94

fred

0.94

hardt

0.91

NER

0.90

ning

0.90

ters

0.86

Activations Density 0.019%

references to the location or person named "Win"

instances of the word "Win" related to various contexts or topics

No Comments

No Known Activations