INDEX

Explanations

numerical values and formatting related to programming errors or debugging information

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

httphttps

-0.79

 tartalomajánló

-0.77

 beginnetje

-0.77

 NSCoder

-0.67

+:+

-0.66

 дописавши

-0.64

uxxxx

-0.63

LookAnd

-0.62

 وتسجيلات

-0.60

 utafitiHapana

-0.59

POSITIVE LOGITS

 setw

0.58

setw

0.50

iomanip

0.47

 padded

0.38

padded

0.35

tournament

0.35

 vertical

0.34

 dynamic

0.34

 textAlign

0.34

padding

0.34

Activations Density 0.018%

numerical values and formatting related to programming errors or debugging information

No Comments

No Known Activations