INDEX

Explanations

personal statements or inner reflections describing oneself

oai_token-act-pair · gpt-3.5-turbo

phrases emphasizing exclusivity or being the only one

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

mares

-0.88

tones

-0.81

cies

-0.76

ieties

-0.73

ulations

-0.71

ships

-0.71

asonry

-0.71

ulas

-0.70

hips

-0.69

yles

-0.69

POSITIVE LOGITS

 smartest

1.28

 bearer

1.17

 happiest

1.17

 beneficiary

1.14

guy

1.11

 rightful

1.08

 embodiment

1.07

 aggress

1.06

 messenger

1.04

 underdog

1.04

Activations Density 0.135%

personal statements or inner reflections describing oneself

phrases emphasizing exclusivity or being the only one

No Comments

No Known Activations