INDEX

Explanations

programming language syntax, particularly related to imports and schema definitions

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

scientific text with code contexts

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

存于互联网档案馆

-0.73

 CURIAM

-0.71

nonatomic

-0.68

TagMode

-0.61

msgTypes

-0.59

CodeAttribute

-0.58

ביוגרפיה

-0.57

ResumeLayout

-0.56

contentLoaded

-0.56

ISPR

-0.56

POSITIVE LOGITS

 Grüsse

0.47

 Grüße

0.45

 asegu

0.44

 contemporánea

0.44

 Könige

0.44

 botellas

0.44

 kautta

0.44

 alrededores

0.43

 gruesa

0.42

 asimismo

0.42

Activations Density 0.005%

programming language syntax, particularly related to imports and schema definitions

scientific text with code contexts

No Comments

No Known Activations