INDEX

Explanations

phrases indicating preferences, roles, and evaluations within contexts

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

club or path

np_acts-logits-general · gemini-2.5-flash-lite

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

Personendaten

-0.53

transQ

-0.52

fjspx

-0.51

Discografia

-0.47

tifact

-0.45

+#+#

-0.45

 Diſ

-0.44

makeConstraints

-0.44

 հղումներ

-0.44

 сер

-0.43

POSITIVE LOGITS

 ThemeData

0.39

鰭

0.36

ExtendWith

0.36

Hozzáférés

0.35

BuildContext

0.35

بوابة

0.34

 boost

0.34

'},

0.33

 clube

0.32

FontOfSize

0.32

Activations Density 0.012%

phrases indicating preferences, roles, and evaluations within contexts

club or path

No Comments

No Known Activations