INDEX

Explanations

specific proper nouns and significant entities related to texts or narratives

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

starting phrases like "from", "one", "who", "so"

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

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

Not in Any Lists

No Comments

Negative Logits

Географиясе

-0.49

hum

-0.34

inputStream

-0.32

 Rule

-0.31

gl

-0.31

SET

-0.31

setOn

-0.31

enum

-0.30

 barrier

-0.29

adaptiveStyles

-0.29

POSITIVE LOGITS

verwijspagina

0.69

 Vikipedi

0.55

ajuku

0.55

CloseOperation

0.54

whom

0.54

interopRequire

0.54

 للمعارف

0.51

ⓧ

0.51

 whom

0.51

AddTagHelper

0.50

Activations Density 0.120%

specific proper nouns and significant entities related to texts or narratives

starting phrases like "from", "one", "who", "so"

No Comments

No Known Activations