INDEX

Explanations

specific brand names or proper nouns, like "Red Bull."

oai_token-act-pair · gpt-3.5-turbo

references to pop culture and entertainment elements

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

etheless

-0.92

quartered

-0.71

 âĵĺ

-0.69

)]

-0.68

rawdownloadcloneembedreportprint

-0.67

nered

-0.65

ciplinary

-0.62

channelAvailability

-0.62

icably

-0.61

itled

-0.60

POSITIVE LOGITS

 Kardash

0.91

 Klux

0.88

zac

0.85

akespe

0.82

 Oprah

0.81

opoly

0.80

 Coke

0.78

George

0.78

 Barbie

0.77

 Beatles

0.77

Activations Density 0.411%

specific brand names or proper nouns, like "Red Bull."

references to pop culture and entertainment elements

No Comments

No Known Activations