INDEX

Explanations

references to names and identifications related to individuals or entities in various contexts

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

mentions of names

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

WS

-0.30

folgen

-0.29

 Door

-0.28

Door

-0.26

 translate

-0.26

 saída

-0.26

 stick

-0.26

 change

-0.25

POSITIVE LOGITS

isContained

0.80

 intStringLen

0.76

SharedCtor

0.71

 iNdEx

0.71

 Administrativna

0.65

RenderAtEndOf

0.62

BASEPATH

0.61

ScopeManager

0.61

SourceChecksum

0.60

twimg

0.59

Activations Density 0.066%

references to names and identifications related to individuals or entities in various contexts

mentions of names

No Comments

No Known Activations