INDEX

Explanations

code-related terminology and operations

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

parsing and exceptions

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

 מר

-0.36

 pungkas

-0.36

odetic

-0.34

ReusableCell

-0.34

bably

-0.33

 méda

-0.33

 φι

-0.32

ゴリー

-0.31

 tamaños

-0.31

POSITIVE LOGITS

parsedMessage

0.64

expandindo

0.61

DataAnnotations

0.59

errHandler

0.57

contentLoaded

0.55

 Wiktionnaire

0.54

 Autorisations

0.54

 kaynağından

0.52

enderror

0.51

 насељу

0.50

Activations Density 0.017%

code-related terminology and operations

parsing and exceptions

No Comments

No Known Activations