INDEX

Explanations

mentions of memory addresses and related terminology

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

address of or width

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

Won

-0.59

 Pauline

-0.52

 العاده

-0.51

 OkHttpClient

-0.50

Won

-0.50

 Vicky

-0.50

Beng

-0.49

Pauline

-0.49

Joe

-0.48

Ke

-0.48

POSITIVE LOGITS

 Address

1.13

 address

1.09

Address

1.02

address

1.01

 ADDRESS

0.94

 getAddress

0.90

ddress

0.88

 Addresses

0.87

 adress

0.87

ADDRESS

0.84

Activations Density 0.020%

mentions of memory addresses and related terminology

address of or width

No Comments

No Known Activations