INDEX

Explanations

intensifiers indicating a high degree of quality or quantity

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

after "very"

np_acts-logits-general · gemini-2.0-flash

intensifier followed by descriptor

np_acts-logits-general · gemini-2.5-flash-lite

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 20-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_20/width_16k/average_l0_71

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 auroit

-0.73

 fometimes

-0.72

 bershka

-0.71

 Jefus

-0.71

 perfons

-0.70

Personendaten

-0.69

 ModelExpression

-0.68

 ſche

-0.68

 Theſe

-0.67

 sonriendo

-0.64

POSITIVE LOGITS

 different

0.68

 effective

0.58

 prevalent

0.57

PDOException

0.56

 successful

0.55

 close

0.55

 easy

0.55

 reliant

0.54

 self

0.53

 hard

0.52

Activations Density 0.340%

intensifiers indicating a high degree of quality or quantity

after "very"

intensifier followed by descriptor

No Comments

No Known Activations