INDEX

Explanations

specific formatting and markup components in code-related texts

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

code and programming terms

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

OGO

-0.42

 Purity

-0.40

 separating

-0.40

ogo

-0.40

 differentiating

-0.39

TagMode

-0.37

 Exclusion

-0.37

 dilu

-0.37

 Seals

-0.37

GEBURTS

-0.37

POSITIVE LOGITS

 NSCoder

0.68

HomeAsUpEnabled

0.63

 AssemblyCompany

0.60

MessageTagHelper

0.57

XtraEditors

0.54

 GenerationType

0.52

AccessorTable

0.52

aimana

0.52

 resourceCulture

0.51

MigrationBuilder

0.49

Activations Density 0.009%

specific formatting and markup components in code-related texts

code and programming terms

No Comments

No Known Activations