INDEX

Explanations

words or phrases related to themed items, places, or events

oai_token-act-pair · gpt-3.5-turbo

instances of themed content across various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

 exert

-0.74

 Figure

-0.65

ox

-0.63

 alone

-0.63

 weak

-0.63

sys

-0.62

 exits

-0.60

li

-0.60

fib

-0.60

 failure

-0.59

POSITIVE LOGITS

themed

3.55

 themed

2.32

inspired

1.98

theme

1.65

focused

1.48

branded

1.42

related

1.29

Theme

1.25

shaped

1.21

 theme

1.11

Activations Density 0.017%

words or phrases related to themed items, places, or events

instances of themed content across various contexts

No Comments

No Known Activations