INDEX

Explanations

images and links within text formats

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

image identifiers

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



-0.64

httphttps

-0.59

帖最后由

-0.57

 <>",

-0.55

OCCURRED

-0.55

StoryboardSegue

-0.54

 intStringLen

-0.53

ftagPool

-0.51

principalTable

-0.50

enderror

-0.47

POSITIVE LOGITS

AnimationsModule

0.48

__(/*!

0.40

 Konf

0.38

 WOLF

0.38

верен

0.38

 ELECTRONIC

0.37

ڠ

0.35

 thọ

0.35

<0x90>

0.34

 Schre

0.34

Activations Density 0.062%

images and links within text formats

image identifiers

No Comments

No Known Activations