INDEX

Explanations

highly repetitive sequences or structures in code

oai_token-act-pair · gpt-4o-mini

Triggered by @bot

code snippets

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

Negative Logits

windowFixed

-0.55

 enfans

-0.54

 CanadaChoose

-0.52

 Reſ

-0.52

adaptiveStyles

-0.50

-------

-0.50

TagMode

-0.49

 faſt

-0.46

TypedDataSet

-0.46

yarnpkg

-0.45

POSITIVE LOGITS

 Melissa

0.45

 Bird

0.44

 Jocelyn

0.43

ValueStyle

0.43

oyd

0.42

Koy

0.42

داری

0.42

ffic

0.41

Hy

0.41

Activations Density 0.003%

highly repetitive sequences or structures in code

code snippets

No Known Activations