INDEX

Explanations

questions that begin with "how," "what," or "where."

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

Questions starting with "How"

np_acts-logits-general · gemini-2.0-flash

quantifiable "how" 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-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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

Хьажоргаш

-0.78

 pleaſure

-0.77

Datuak

-0.68

 purpoſe

-0.65

GEBURTSDATUM

-0.63

 stiefel

-0.63

Portail

-0.61

MigrationBuilder

-0.61



-0.60

Portale

-0.58

POSITIVE LOGITS

 exactly

0.68

 much

0.63

 aware

0.56

 typelib

0.55

 long

0.55

 well

0.52

 many

0.52

 best

0.52

 else

0.50

 nhiêu

0.50

Activations Density 0.078%

questions that begin with "how," "what," or "where."

Questions starting with "How"

quantifiable "how" questions

No Comments

No Known Activations