INDEX

Explanations

past tense verbs and phrases indicating prior actions or conditions

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

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Inscrivez

-0.48

guiente

-0.44

 obecnie

-0.41

AnimationsModule

-0.41

 nästa

-0.40

 soeur

-0.39

lorette

-0.39

 ainfi

-0.39

värr

-0.38

 atualmente

-0.38

POSITIVE LOGITS

 था

0.67

was

0.66

وكان

0.64

 meant

0.62

 était

0.61

 היה

0.61

と思っていた

0.60

Twas

0.60

 Había

0.60

Was

0.59

Activations Density 0.004%

past tense verbs and phrases indicating prior actions or conditions

No Comments

No Known Activations