INDEX

Explanations

nested structures or hierarchical patterns in data representations

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

indices and variables

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

RenderAtEndOf

-0.62

AddTagHelper

-0.60

BibitemShut

-0.58

AndroidJUnit

-0.54

 pinulongan

-0.54

 kasarigan

-0.54

RegistryLite

-0.53

fromCharCode

-0.52

+#+

-0.50

 wireType

-0.50

POSITIVE LOGITS

 between

0.42

="../../

0.42

://

0.38

="../

0.37

Between

0.37

between

0.37

 Between

0.36

={`/

0.35

 BETWEEN

0.35

him

0.35

Activations Density 0.051%

nested structures or hierarchical patterns in data representations

indices and variables

No Comments

No Known Activations