INDEX

Explanations

negations and forms of denial

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

negative contractions

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

 ujednoznacz

-0.67

-0.66

 Chwiliwch

-0.58

存于互联网档案馆

-0.56

″]

-0.56

printStackTrace

-0.51

TargetException

-0.50

angsaan

-0.49

EIF

-0.48

ArgsConstructor

-0.46

POSITIVE LOGITS

wouldn

0.90

 wouldn

0.88

 wasn

0.84

 doesn

0.83

<bos>

0.83

Wouldn

0.83

 didn

0.83

 weren

0.81

wasn

0.79

 aren

0.79

Activations Density 0.057%

negations and forms of denial

negative contractions

No Comments

No Known Activations