INDEX

Explanations

medical terms and location names related to specific countries

oai_token-act-pair · gpt-3.5-turbo

proper nouns related to places and institutions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

xp

-0.59

pak

-0.59

 tick

-0.57

 clicks

-0.57

 stakes

-0.57

 caps

-0.57

 SPACE

-0.57

 crop

-0.56

 haste

-0.55

 space

-0.55

POSITIVE LOGITS

elaide

0.89

odore

0.88

xon

0.84

west

0.83

oldemort

0.77

agall

0.76

rimination

0.75

endment

0.75

bnb

0.75

alyst

0.75

Activations Density 0.205%

medical terms and location names related to specific countries

proper nouns related to places and institutions

No Comments

No Known Activations