INDEX

Explanations

mentions of the word "penguins"

oai_token-act-pair · gpt-3.5-turbo

references to penguins

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

WORK

-0.82

ysis

-0.80

puter

-0.73

phas

-0.71

usted

-0.66

¿½

-0.66

 ILCS

-0.65

neau

-0.65

parts

-0.65

lder

-0.64

POSITIVE LOGITS

insula

0.99

 Penguins

0.86

 pengu

0.82

aukee

0.77

atoon

0.77

engu

0.74

keye

0.72

 Pengu

0.72

unia

0.71

 Hots

0.67

Activations Density 0.025%

mentions of the word "penguins"

references to penguins

No Comments

No Known Activations