INDEX

Explanations

phrases describing physical location and actions related to concealing objects

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

New Auto-Interp

Configuration

jbloom/Gemma-2b-IT-Residual-Stream-SAEs/gemma_2b_it_blocks.12.hook_resid_post_16384

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

HuggingFaceFW/fineweb

Features

16,384

Data Type

float32

Hook Name

blocks.12.hook_resid_post

Hook Layer

Architecture

standard

Context Size

1,024

Dataset

Skylion007/openwebtext

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 succede

-0.68

uhr

-0.55

 antici

-0.54

 impon

-0.54

 Minangkabau

-0.54

 conclud

-0.53

 rheumat

-0.53

 ineffec

-0.53

 quí

-0.52

 horrend

-0.51

POSITIVE LOGITS

 hidden

0.67

 hiding

0.55

hidden

0.54

 buried

0.53

 Hidden

0.51

 beneath

0.51

HIDDEN

0.50

 löyty

0.50

 storage

0.50

Hidden

0.49

Activations Density 0.303%

phrases describing physical location and actions related to concealing objects

No Comments

No Known Activations