INDEX

Explanations

uncommon or hidden things or entities, possibly with a secretive or mysterious context

oai_token-act-pair · gpt-3.5-turbo

references to the concept of being unknown or undiscovered

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

owell

-0.75

 Suff

-0.71

ulse

-0.70

SQ

-0.68

EE

-0.66

 affirmative

-0.65

oline

-0.64

 battery

-0.64

 shudder

-0.62

 maid

-0.61

POSITIVE LOGITS

nown

1.27

friends

0.85

vous

0.84

sta

0.81

cia

0.78

gnu

0.78

=-=-

0.77

tesy

0.75

 Skydragon

0.75

gang

0.74

Activations Density 0.033%

uncommon or hidden things or entities, possibly with a secretive or mysterious context

references to the concept of being unknown or undiscovered

No Comments

No Known Activations