INDEX

Explanations

proper nouns related to technology and individuals in news or entertainment

oai_token-act-pair · gpt-3.5-turbo

references to organizations and notable individuals

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

etheless

-0.98

nce

-0.70

GoldMagikarp

-0.69

utory

-0.66

 Cosponsors

-0.64

}{

-0.64

":"","

-0.63

general

-0.63

norm

-0.63

ends

-0.63

POSITIVE LOGITS

 intact

1.00

 hooked

0.90

 installed

0.88

 handy

0.85

 locked

0.82

 ready

0.82

 attached

0.82

 removed

0.81

 hostage

0.81

 gone

0.78

Activations Density 0.617%

proper nouns related to technology and individuals in news or entertainment

references to organizations and notable individuals

No Comments

No Known Activations