INDEX

Explanations

programming-related syntax and structures

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

fn, impl, pub fn

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

//

-0.31

DotNetBar

-0.29

 preto

-0.29

 deshalb

-0.28

ile

-0.27

 Egipto

-0.26

Egypt

-0.26

frac

-0.25

 sayang

-0.25

POSITIVE LOGITS

 AssemblyVersion

0.86

httphttps

0.73

zzleHttp

0.73

DockStyle

0.69

 rhestr

0.69

 nahilalakip

0.66

iſchen

0.63

InputBorder

0.62

 '\\;'

0.61

๣

0.60

Activations Density 0.044%

programming-related syntax and structures

fn, impl, pub fn

No Comments

No Known Activations