INDEX

Explanations

references to "light" or "lightweight" concepts and objects

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 11-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_11/width_16k/average_l0_80

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.11.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

 propOrder

-1.17

 оригіналу

-1.03

ganu

-1.02

quelize

-0.98

__":

-0.97

AndEndTag

-0.95

 تانيه

-0.94

 blessés

-0.94

 bezeichneter

-0.93

 ProtoMessage

-0.93

POSITIVE LOGITS

 Light

1.56

Light

1.49

 LIGHT

1.48

light

1.44

 light

1.38

LIGHT

1.35

 Lightfoot

1.12

lights

1.09

 Lights

1.06

 ligh

1.04

Activations Density 0.080%

references to "light" or "lightweight" concepts and objects

No Comments

No Known Activations