INDEX

Explanations

phrases indicating exclusivity or limitations

oai_token-act-pair · gpt-4o-mini

Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 13-gemmascope-res-65k

Configuration

google/gemma-scope-2b-pt-res/layer_13/width_65k/average_l0_75

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

65,536

Data Type

float32

Hook Name

blocks.13.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

IFrame

Link

Negative Logits

\{\\

-0.75

 Roskov

-0.70

TagMode

-0.65

WriteTagHelper

-0.65

 indé

-0.63

GETHER

-0.61

 inspiré

-0.60

Schedulers

-0.60

 AssemblyCulture

-0.59

 russa

-0.59

POSITIVE LOGITS

 only

0.74

 really

0.70

 thing

0.69

 downside

0.69

 exception

0.68

 exceptions

0.67

 best

0.62

 ever

0.60

 reason

0.60

 unknownFields

0.59

Activations Density 0.025%

phrases indicating exclusivity or limitations

No Known Activations