INDEX

Explanations

concepts related to artificial intelligence and intelligent systems

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

human or intelligent action

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

mentions of intelligence, intelligent agents, perception/awareness, or cognitive capabilities (e.g., words about intelligent systems, perceiving, acting, or being aware).

oai_token-act-pair · gpt-5-mini Triggered by @vetterc0

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

 يتيمه

-0.47

elemField

-0.45

 NSCoder

-0.42

Hozzáférés

-0.41

PCell

-0.40

Carriera

-0.40

Personendaten

-0.40

❋

-0.39

pone

-0.39

 للمعارف

-0.38

POSITIVE LOGITS

 autonomía

0.45

 Reſ

0.45

 humano

0.44

 humanos

0.43

 humanas

0.42

 menschliche

0.42

 Houſe

0.42

 humanities

0.42

 umane

0.41

 ſen

0.40

Activations Density 0.024%

concepts related to artificial intelligence and intelligent systems

human or intelligent action

mentions of intelligence, intelligent agents, perception/awareness, or cognitive capabilities (e.g., words about intelligent systems, perceiving, acting, or being aware).

No Comments

No Known Activations

concepts related to artificial intelligence and intelligent systems

human or intelligent action

mentions of intelligence, intelligent agents, perception/awareness, or cognitive capabilities (e.g., words about intelligent systems, perceiving, acting, or being aware).

No Comments

No Known Activations