INDEX

Explanations

words that reference quantities or numbers

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 15-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_15/width_16k/average_l0_78

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.15.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

ſelves

-0.82

 BRARY

-0.77

ſelf

-0.74

 iſt

-0.73

MessageOf

-0.72

 Priv

-0.71

 Spital

-0.71

InstrumentedTest

-0.69

GEBUR

-0.68

neſs

-0.67

POSITIVE LOGITS

newArrayList

0.51

...

0.47

 okuyayım

0.46

 vapore

0.45

MessageTagHelper

0.45

to

0.44

 çıkar

0.44

 liburan

0.44

Activations Density 0.080%

words that reference quantities or numbers

No Comments

No Known Activations