INDEX

Explanations

programming-related keywords in a code-like structure

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

code comments and citations

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

ChildScrollView

-0.36

 Infórmanos

-0.33

нома

-0.33

 critic

-0.32

 greetings

-0.32

soi

-0.31

そこで

-0.31

RuleContext

-0.30

Normalized

-0.30

 process

-0.30

POSITIVE LOGITS

 ujednoznacz

0.68

 disambiguazione

0.61

 ModelExpression

0.60

 intptr

0.54

retudo

0.54

 referrerpolicy

0.54

قایناق‌لار

0.54

点此举报

0.52

 gynhyrchwyd

0.52

enderror

0.52

Activations Density 0.003%

programming-related keywords in a code-like structure

code comments and citations

No Comments

No Known Activations