INDEX

Explanations

terms related to chemical structures or compounds, specifically those that might signify functional groups

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

hyper-responsiveness or mechanical

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

SequentialGroup

-0.84

 feroit

-0.77

 itſelf

-0.75

 plufieurs

-0.75

 ainfi

-0.73

 ſont

-0.71

 ſich

-0.71

 CreateTagHelper

-0.69

 الحره

-0.69

 myſelf

-0.69

POSITIVE LOGITS

0.86

0.67

–

0.56

-,

0.56

-'

0.54

‐

0.52

or

0.51

and

0.49

-(

0.49

Activations Density 0.054%

terms related to chemical structures or compounds, specifically those that might signify functional groups

hyper-responsiveness or mechanical

No Comments

No Known Activations