INDEX

Explanations

specific object identifiers and operational terms in code and technical descriptions

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

bearing attached to

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

 Wikimedijinoj

-0.45

Spoljašnje

-0.43

WriteTagHelper

-0.41

 Савезне

-0.38

 kaarangay

-0.35

Források

-0.35

migrationBuilder

-0.35

sidemargin

-0.35

 Mejora

-0.35

 StatelessWidget

-0.34

POSITIVE LOGITS

assies

0.56

Бахар

0.51

yer

0.50

RTGC

0.50

domada

0.49

werbe

0.49

 насељу

0.47

transQ

0.47

udent

0.46

↹

0.46

Activations Density 0.169%

specific object identifiers and operational terms in code and technical descriptions

bearing attached to

No Comments

No Known Activations