INDEX

Explanations

the word "Magn" or "Magnus"

oai_token-act-pair · gpt-3.5-turbo

references to "Magnitsky" and related topics involving legal issues and human rights

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

resp

-0.86

CHAT

-0.84

rior

-0.82

viously

-0.79

clamation

-0.78

OPLE

-0.76

vious

-0.71

aper

-0.70

$$

-0.69

FAULT

-0.66

POSITIVE LOGITS

ificent

1.23

olia

0.95

achev

0.89

animous

0.85

iment

0.82

itude

0.80

 cones

0.80

ertodd

0.78

wig

0.77

 glim

0.77

Activations Density 0.022%

the word "Magn" or "Magnus"

references to "Magnitsky" and related topics involving legal issues and human rights

No Comments

No Known Activations