INDEX

Explanations

phrases that indicate appearances or assumptions

oai_token-act-pair · gpt-3.5-turbo

phrases that describe appearances or observations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

called

-0.70

ãĤ¦ãĤ¹

-0.65

imore

-0.64

cart

-0.62

eting

-0.58

cest

-0.58

anwhile

-0.58

DAQ

-0.57

TED

-0.57

opens

-0.57

POSITIVE LOGITS

 violate

0.97

 contradict

0.93

 condone

0.89

 defy

0.86

be

0.85

 indicate

0.85

 soften

0.83

 confirm

0.83

 imitate

0.83

 originate

0.81

Activations Density 0.061%

phrases that indicate appearances or assumptions

phrases that describe appearances or observations

No Comments

No Known Activations