INDEX

Explanations

words related to observing or looking at something

oai_token-act-pair · gpt-3.5-turbo

instances of the word "watch" 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 @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

esan

-0.88

ãĥ´

-0.83

enture

-0.74

interstitial

-0.69

ItemTracker

-0.68

alty

-0.68

eno

-0.67

wealth

-0.67

hesion

-0.66

afort

-0.66

POSITIVE LOGITS

dogs

1.11

 unfold

0.98

tv

0.98

dog

0.98

TV

0.98

 helpless

0.96

 videos

0.95

 television

0.95

 movies

0.92

 Watching

0.89

Activations Density 0.066%

words related to observing or looking at something

instances of the word "watch" and its variations

No Comments

No Known Activations