INDEX

Explanations

specific keywords and symbols related to programming or code structures

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

conditional, given, text

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

 ""],

-0.41

?”.

-0.32

 ""),

-0.31

enschappe

-0.30

 uitz

-0.30

."],

-0.30

 spoko

-0.29

 ''),

-0.29

;;)

-0.29

"]),

-0.29

POSITIVE LOGITS

RTCF

0.57

yyr

0.57

teaser

0.57

 ब्रेकडाउन

0.56

 Strategy

0.56

########.

0.56

ſammen

0.56

 Triggers

0.55

prepareStatement

0.55

 ويكيپ

0.55

Activations Density 0.002%

specific keywords and symbols related to programming or code structures

conditional, given, text

No Comments

No Known Activations