INDEX

Explanations

text related to childhood experiences or memories

oai_token-act-pair · gpt-3.5-turbo

mentions of the word "kid."

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

veyard

-0.82

aukee

-0.76

anwhile

-0.75

 heaviest

-0.72

etting

-0.71

election

-0.71

aeda

-0.68

 stocking

-0.67

 plurality

-0.66

ĺħ

-0.66

POSITIVE LOGITS

kid

1.12

kids

0.98

ishly

0.88

aign

0.88

 kids

0.88

 Doodle

0.88

 Icar

0.84

boys

0.83

Kid

0.80

 Gohan

0.80

Activations Density 0.005%

text related to childhood experiences or memories

mentions of the word "kid."

No Comments

No Known Activations