INDEX

Explanations

numerical data related to measurements or costs

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

average, typical, and size estimations

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

harapkan

-0.58

WriteBarrier

-0.55

 dezelve

-0.52

 zoude

-0.51

LookAnd

-0.49

AsUp

-0.47

DockStyle

-0.46

 snippetHide

-0.46

antaranya

-0.45

@"/

-0.45

POSITIVE LOGITS

UrlResolution

0.46

 average

0.41

 typical

0.38

 Typical

0.38

 Average

0.37

igno

0.36

 Vina

0.35

 assume

0.34

 opinion

0.34

Average

0.34

Activations Density 0.052%

numerical data related to measurements or costs

average, typical, and size estimations

No Comments

No Known Activations