INDEX

Explanations

references to warfare and conflict-related themes

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 20-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_20/width_16k/average_l0_71

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

+#+#

-0.71

WriteLiteral

-0.58

 Fick

-0.57

 sanitaires

-0.55

Bands

-0.52

 Bens

-0.50

 météo

-0.49

Permit

-0.49



-0.49

Kontrola

-0.49

POSITIVE LOGITS

 Transformers

0.85

 TRANSFORM

0.83

 Optimus

0.82

 Auto

0.80

gatron

0.77

Transformers

0.76

 transform

0.74

transform

0.74

TRANSFORM

0.74

 Transformer

0.72

Activations Density 0.180%

references to warfare and conflict-related themes

No Comments

No Known Activations