INDEX

Explanations

elements related to programming or technical specifications

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

performing with

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

'{@

-0.37

戬

-0.35

laması

-0.34

 éve

-0.33

Vidite

-0.33

nsf

-0.32

ouching

-0.32

kool

-0.31

hota

-0.31

 wodurch

-0.30

POSITIVE LOGITS

IndentedString

0.60

 EconPapers

0.60

errHandler

0.59

SourceChecksum

0.53

 cherchés

0.51

createCanvas

0.51

DockStyle

0.51

@[+][

0.50

 CreateTagHelper

0.49

 Meksiku

0.48

Activations Density 0.066%

elements related to programming or technical specifications

performing with

No Comments

No Known Activations