INDEX

Explanations

highly significant numerical values or data-related information

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 4-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_4/width_16k/average_l0_124

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.4.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

WriteLiteral

-0.68

endpush

-0.62

lysses

-0.62

 Waray

-0.61

DebuggerNonUser

-0.61

CodeAttribute

-0.58

 exponer

-0.58

 România

-0.57

moveToFirst

-0.56

vertre

-0.56

POSITIVE LOGITS

DataAnnotations

0.67

 propOrder

0.66

ModelAdmin

0.64

 poffible

0.63

annica

0.61

 ninety

0.60

 arrière

0.59

 sûr

0.59

ronesia

0.59

cemic

0.58

Activations Density 0.319%

highly significant numerical values or data-related information

No Comments

No Known Activations