INDEX

Explanations

connections and relationships between entities or experiences

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

defining abstract nouns

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

 kuitenkin

-0.51

 nevertheless

-0.47

السكان

-0.45

 aapt

-0.40

UnusedPrivate

-0.40

 AssemblyCulture

-0.40

TabIndex

-0.40

期刊论文

-0.39

 bowiem

-0.39

 nonetheless

-0.39

POSITIVE LOGITS

+#+#

0.67

mybatisplus

0.48

braio

0.42

DeleteBehavior

0.39

specially

0.39

CreateMap

0.38

EDEFAULT

0.38

 ration

0.37

Ї

0.36

SNAP

0.36

Activations Density 0.070%

connections and relationships between entities or experiences

defining abstract nouns

No Comments

No Known Activations