INDEX

Explanations

specific actions or behaviors related to decision-making and evaluation processes

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

actions and states

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

 Absicht

-0.38

 lunak

-0.38

 chví

-0.37

ạp

-0.36

 Económica

-0.34

 oreja

-0.34

 στιγ

-0.34

 Hinsicht

-0.33

 šed

-0.33

 ciepła

-0.33

POSITIVE LOGITS

BagLayout

0.57

jsxFileName

0.56

 kaarangay

0.55

adaptiveStyles

0.52

 pinulongan

0.50

ImageContext

0.50

 nakalista

0.50

 ddelweddau

0.50

ंदीखरीदारी

0.49

 PeEnEo

0.49

Activations Density 0.079%

specific actions or behaviors related to decision-making and evaluation processes

actions and states

No Comments

No Known Activations