INDEX

Explanations

elements or symbols typically associated with coding or mathematical notation

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

Hexadecimal representations, often starting with "0x"

np_acts-logits-general · gemini-2.0-flash

hexadecimal literals

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

 juſ

-0.84

 betweenstory

-0.78

DockStyle

-0.76

NameInMap

-0.75

LookAnd

-0.75

 pleaſure

-0.74

 ſta

-0.74

 ſte

-0.72

 deſt

-0.71

исленность

-0.70

POSITIVE LOGITS

""".

0.57

withIdentifier

0.47

)}$.

0.45

 geworden

0.45

uitton

0.43

/");

0.43

",$

0.42

廓

0.41

[toxicity=0]

0.41

Activations Density 0.266%

elements or symbols typically associated with coding or mathematical notation

Hexadecimal representations, often starting with "0x"

hexadecimal literals

No Comments

No Known Activations