INDEX

Explanations

connective words and phrases indicating relationships or additions

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

what and why questions

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-65k

Configuration

google/gemma-scope-2b-pt-res/layer_20/width_65k/average_l0_114

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

65,536

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

 GenerationType

-0.55

 fili

-0.53

 parap

-0.52

AutoField

-0.51

таратура

-0.51

 DataGridView

-0.51

商品説明

-0.49

 ujednoznacz

-0.49

 onPostExecute

-0.48

czuk

-0.47

POSITIVE LOGITS

 what

0.92

 mitä

0.76

what

0.76

 aquilo

0.70

 ciò

0.67

 whom

0.63

 aquello

0.62

quilo

0.62

 quello

0.61

 ceea

0.59

Activations Density 0.043%

connective words and phrases indicating relationships or additions

what and why questions

No Comments

No Known Activations