INDEX

Explanations

words related to options, choices, variations, and different configurations

oai_token-act-pair · gpt-3.5-turbo

references to variations or differences in context or characteristics

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

aught

-0.79

ulner

-0.76

========

-0.71

average

-0.64

elfare

-0.60

+++

-0.60

 newcomers

-0.60

olicited

-0.59

apego

-0.59

raised

-0.57

POSITIVE LOGITS

 places

1.34

 directions

1.32

 contexts

1.31

 locations

1.25

 direction

1.19

 guise

1.18

 manner

1.18

 vicinity

1.15

 fashion

1.10

 manners

1.10

Activations Density 0.185%

words related to options, choices, variations, and different configurations

references to variations or differences in context or characteristics

No Comments

No Known Activations