INDEX

Explanations

references to specific biological or scientific concepts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 17-gemmascope-res-65k

Configuration

google/gemma-scope-2b-pt-res/layer_17/width_65k/average_l0_125

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

65,536

Data Type

float32

Hook Name

blocks.17.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 loss

-0.44

chengladbach

-0.43

 gere

-0.43

 ανα

-0.41

pic

-0.41

zielle

-0.41

 gefähr

-0.41

loss

-0.40

ciano

-0.40

 Loss

-0.40

POSITIVE LOGITS

 ſtill

0.73

 myſelf

0.72

 propOrder

0.71

BeginContext

0.70

astify

0.68

 saites

0.68

DotNetBar

0.67

ScopeManager

0.67

IndentedString

0.66

 bukan

0.65

Activations Density 0.312%

references to specific biological or scientific concepts

No Comments

No Known Activations