INDEX

Explanations

instances of object or variable names in programming or data descriptions

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

Weizmann Institute

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

########.

-0.59

JspWriter

-0.50

findpost

-0.49

rzost

-0.48

Rhestr

-0.48

 snippetHide

-0.46

nloa

-0.46

 surla

-0.44

RenderAtEndOf

-0.41

COUVER

-0.40

POSITIVE LOGITS

name

0.59

DisplayName

0.54

 AssemblyCompany

0.51

LabelTagHelper

0.49

 name

0.47

NAME

0.45

 bílá

0.44

🔰

0.43

ylabel

0.42

EndInit

0.42

Activations Density 0.000%

instances of object or variable names in programming or data descriptions

Weizmann Institute

No Comments

No Known Activations