INDEX

Explanations

technical terms and code-related syntax elements

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

hit, impact, or collisions

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

 gethan

-0.39

omitempty

-0.39

 QtWidgets

-0.36

gemeester

-0.36

BASEPATH

-0.35

 selectAll

-0.35

AccessFile

-0.34

lwjgl

-0.34

 especí

-0.34

 hảo

-0.34

POSITIVE LOGITS

 collision

0.52

+:+

0.51

pecabe

0.48

 collisions

0.46

Cyfeiriadau

0.46

enderror

0.45

衝突

0.43

MessageTagHelper

0.43

collision

0.42

++

0.42

Activations Density 0.003%

technical terms and code-related syntax elements

hit, impact, or collisions

No Comments

No Known Activations