INDEX

Explanations

phrases that indicate connections to various institutions or standards

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

strategies and quality

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

Not in Any Lists

No Comments

Negative Logits

CROSSTALK

-0.41

 Audiodateien

-0.40

 Assistant

-0.40

Cuántos

-0.39

OfWeek

-0.37

Assistant

-0.36

orElse

-0.36

AndEndTag

-0.36

Weiterlesen

-0.35

 César

-0.35

POSITIVE LOGITS

contentLoaded

0.58

 quality

0.50

quality

0.49

GTCX

0.48

 Qualität

0.48

Quality

0.48

HideFlags

0.47

دانشنامهٔ

0.46

 carefully

0.46

createStatement

0.45

Activations Density 0.015%

phrases that indicate connections to various institutions or standards

strategies and quality

No Comments

No Known Activations