INDEX

Explanations

references to programming or coding terminology

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

:✨

-0.47

 sanitarias

-0.46

 scattata

-0.46

 verdaderamente

-0.44

 prefeitura

-0.44

 knji

-0.44

+#+#

-0.43

 namanya

-0.43

 macetas

-0.42

extAlignment

-0.42

POSITIVE LOGITS

 Valve

0.69

 Dota

0.68

Dota

0.61

 Steam

0.60

Valve

0.58

dota

0.56

 aureus

0.53

 antimony

0.52

 Valves

0.52

steamcommunity

0.51

Activations Density 0.008%

references to programming or coding terminology

No Comments

No Known Activations