INDEX

Explanations

proper nouns, specifically names of people and locations

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

names followed by said

np_acts-logits-general · gemini-2.5-flash-lite

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 9-gemmascope-res-131k

Configuration

google/gemma-scope-9b-it-res/layer_9/width_131k/average_l0_121

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.9.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

 kasarigan

-0.77

oa̍t

-0.63

 Roskov

-0.63

 tartalomajánló

-0.61

 AssemblyProduct

-0.59

liothèque

-0.58

 Италијани

-0.57

 Meksiku

-0.56

bnf

-0.56

بوابة

-0.55

POSITIVE LOGITS

 said

0.46

 guess

0.46

 Guess

0.40

Guess

0.36

 reminded

0.36

guess

0.35

 told

0.35

 replied

0.35

 whit

0.34

 thought

0.33

Activations Density 0.018%

proper nouns, specifically names of people and locations

names followed by said

No Comments

No Known Activations