INDEX

Explanations

programming-related identifiers and their associated values

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

register and function calls

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

 NSCoder

-0.86

 estekak

-0.86

VYMaps

-0.84

 виправивши

-0.83

تقاوى

-0.83



-0.82

OGND

-0.82

 ब्रेकडाउन

-0.81

expandindo

-0.79

SharedDtor

-0.77

POSITIVE LOGITS

has

0.33

sphase

0.30

 atas

0.29

HomeAsUp

0.29

type

0.29

 Kasus

0.29

static

0.28

 Partei

0.28

ito

0.28

base

0.27

Activations Density 0.013%

programming-related identifiers and their associated values

register and function calls

No Comments

No Known Activations