INDEX

Explanations

the word "arm" with different affixes or in compounds

oai_token-act-pair · gpt-3.5-turbo

mentions of "karmic" or related terms

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

 Vide

-0.73

 porous

-0.71

ween

-0.70

 moot

-0.65

 Atlantis

-0.62

rhy

-0.61

 exped

-0.59

 elusive

-0.59

 Gutenberg

-0.58

ĸļ

-0.58

POSITIVE LOGITS

ageddon

1.37

ament

1.23

aments

1.09

arm

1.02

chair

0.98

ory

0.98

achine

0.97

illary

0.95

arms

0.93

esan

0.92

Activations Density 0.006%

the word "arm" with different affixes or in compounds

mentions of "karmic" or related terms

No Comments

No Known Activations