INDEX

Explanations

references to small objects or entities

oai_token-act-pair · gpt-3.5-turbo

instances of the word "little" or variations related to smallness

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

iership

-0.79

ï¸

-0.75

sem

-0.75

idents

-0.74

restling

-0.74

orthy

-0.73

inction

-0.72

ontent

-0.71

worthiness

-0.71

intent

-0.71

POSITIVE LOGITS

 girl

1.10

boy

1.06

 helper

0.99

 sister

0.98

bit

0.98

 brother

0.97

guy

0.92

kid

0.89

 boys

0.87

 girls

0.86

Activations Density 0.032%

references to small objects or entities

instances of the word "little" or variations related to smallness

No Comments

No Known Activations