INDEX

Explanations

the word "most" in various contexts, indicating a focus on superlative or prevalent qualities

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 25-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_25/width_16k/average_l0_116

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.25.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

principalTable

-0.60

 tantôt

-0.53

 perpétu

-0.52

terraform

-0.52

jegy

-0.51

 snabb

-0.50

 milliers

-0.49

بعض

-0.49

 vaikka

-0.49

ناك

-0.48

POSITIVE LOGITS

 importantly

1.56

 notably

1.18

 likely

1.16

 Importantly

1.01

likely

0.92

 Likely

0.89

 prominently

0.83

 recently

0.83

 assuredly

0.81

 interestingly

0.81

Activations Density 0.132%

the word "most" in various contexts, indicating a focus on superlative or prevalent qualities

No Comments

No Known Activations