INDEX

Explanations

phrases and terms associated with uncertainty or ambiguity

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

clear

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

ⓧ

-0.51

FontOfSize

-0.49

operazione

-0.43

CPF

-0.43

 minimalist

-0.42

こうして

-0.41

 ModelRenderer

-0.41

ностях

-0.40

glLoadIdentity

-0.39

san

-0.39

POSITIVE LOGITS

 unclear

0.60

 whether

0.59

 unknown

0.59

 فريبيس

0.55

 Unknown

0.53

unknown

0.50

 UNKNOWN

0.48

unk

0.47

 inconnu

0.46

 nahilalakip

0.45

Activations Density 0.036%

phrases and terms associated with uncertainty or ambiguity

clear

No Comments

No Known Activations