INDEX

Explanations

instances of clarification or definition in a discussion

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

continue to understand

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

ۜ

-0.54

forRoot

-0.50

WindowConstants

-0.50

Ƚ

-0.45

mergeFrom

-0.43

MemoryWarning

-0.43

balleur

-0.43

 snippetHide

-0.43

makeText

-0.42

Microkernel

-0.42

POSITIVE LOGITS

 Blockly

0.41

MILLISECONDS

0.40

 anyway

0.40

 doigt

0.40

DrawerToggle

0.40

anyway

0.39

 explications

0.39

 clarify

0.39

 plutôt

0.38

drawiam

0.38

Activations Density 0.049%

instances of clarification or definition in a discussion

continue to understand

No Comments

No Known Activations