INDEX

Explanations

programming-related constructs and syntax

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

code block identifiers

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

tagHelperRunner

-0.75

transQ

-0.72

AndroidJUnit

-0.69

 resourceCulture

-0.69

RegressionTest

-0.68

tinyos

-0.67

 TestBed

-0.66

addContainerGap

-0.64

 betweenstory

-0.59

Lyd

-0.58

POSITIVE LOGITS

期刊论文

0.54

 uppgifter

0.40

 världen

0.40

Handlung

0.40

 Relaciones

0.39

 stället

0.39

 Sprache

0.36

Népesség

0.35

agaimana

0.35

 návr

0.35

Activations Density 0.009%

programming-related constructs and syntax

code block identifiers

No Comments

No Known Activations