INDEX

Explanations

phrases indicating a specific time period in relation to an event

oai_token-act-pair · gpt-3.5-turbo

references to previous events or conditions

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

hack

-0.66

 elves

-0.66

tower

-0.65

aden

-0.65

ickle

-0.64

asp

-0.64

girls

-0.63

%%%%

-0.63

 Pistons

-0.63

Hub

-0.63

POSITIVE LOGITS

itiz

1.03

itized

0.95

ebin

0.90

etheless

0.90

emort

0.87

ĨĴ

0.84

requisites

0.83

 nomine

0.82

cies

0.79

icip

0.78

Activations Density 0.009%

phrases indicating a specific time period in relation to an event

references to previous events or conditions

No Comments

No Known Activations