INDEX

Explanations

expressions or phrases that indicate parentheses or other forms of nested punctuation

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

IFrame

Link

Negative Logits

ftagPool

-0.53

bootstrapcdn

-0.46

argout

-0.45

 ſur

-0.43

WithMany

-0.42

 Perſ

-0.42

 ſch

-0.41

진

-0.41

 saving

-0.40

 cref

-0.40

POSITIVE LOGITS

GIVEREF

0.57

 EconPapers

0.51

FunctionFlags

0.49

 ferons

0.44

 trám

0.43

WriteTagHelper

0.43

ReusableCell

0.43

astéro

0.42

SOUNDBITE

0.41

Дереккөздер

0.41

Activations Density 0.127%

expressions or phrases that indicate parentheses or other forms of nested punctuation

No Known Activations