INDEX

Explanations

phrases related to ideas, beliefs, opinions, and positions

oai_token-act-pair · gpt-3.5-turbo

statements of necessity or importance regarding a topic

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

iates

-0.76

ragon

-0.66

vet

-0.65

ravel

-0.65

angering

-0.63

illon

-0.62

 happ

-0.61

quer

-0.61

ords

-0.61

 Tanz

-0.60

POSITIVE LOGITS

 namely

0.86

ãĤ¤ãĥĪ

0.75

Hey

0.71

 disclaimer

0.71

"'

0.70

Hey

0.68

andum

0.66

 falsehood

0.66

that

0.66

 disbelief

0.66

Activations Density 0.471%

phrases related to ideas, beliefs, opinions, and positions

statements of necessity or importance regarding a topic

No Comments

No Known Activations