INDEX

Explanations

references to potential dangers or threats in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 1-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_1/width_16k/average_l0_69

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.1.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

rungsseite

-0.59

⤹

-0.52

enumii

-0.52

styleable

-0.47

TestTools

-0.47

 HttpNotFound

-0.47

 Vaux

-0.46

URLException

-0.45

featureID

-0.44

calaureate

-0.44

POSITIVE LOGITS

 threat

0.73

 Threat

0.68

Threat

0.60

 Threats

0.59

 threats

0.56

 default

0.55

 threatening

0.53

Default

0.53

 Default

0.52

threat

0.51

Activations Density 0.274%

references to potential dangers or threats in various contexts

No Comments

No Known Activations