INDEX

Explanations

numerical information and specific identifiers in text

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 20-gemmascope-res-131k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_131k/average_l0_81

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.20.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

balleur

-0.61

 ComVisible

-0.54

 Infórmanos

-0.51

StructEnd

-0.47

 męski

-0.47

reportWebVitals

-0.46

HandlerContext

-0.46

 nonUne

-0.46

RTSN

-0.45

 bParam

-0.43

POSITIVE LOGITS

点此举报

0.40

Pener

0.38

 unresponsive

0.38

 Thank

0.38

gra

0.37

 testimonials

0.37

Gra

0.36

 unauthorized

0.36

 Assembly

0.36

 testimonies

0.36

Activations Density 0.018%

numerical information and specific identifiers in text

No Comments

No Known Activations