INDEX

Explanations

the phrase "type of" followed by a noun or noun phrase

oai_token-act-pair · gpt-3.5-turbo

references to various classifications or categories

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

å§«

-0.76

pload

-0.70

romeda

-0.70

 Rings

-0.69

 Bots

-0.69

olulu

-0.66

 Mald

-0.65

 Kenya

-0.64

IRO

-0.64

 Liberties

-0.63

POSITIVE LOGITS

face

1.26

faces

1.19

etter

1.03

etting

0.97

casting

0.93

ahead

0.84

classes

0.80

olerance

0.78

lander

0.77

cast

0.75

Activations Density 0.021%

the phrase "type of" followed by a noun or noun phrase

references to various classifications or categories

No Comments

No Known Activations