INDEX

Explanations

references to the concept of self

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

references to identity or self-concept

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

self- compounds

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 20-gemmascope-res-65k

Configuration

google/gemma-scope-2b-pt-res/layer_20/width_65k/average_l0_114

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

65,536

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

 "..\..\

-0.83

 "..\..\..\

-0.81

 дописавши

-0.79

Personensuche

-0.78

 الحره

-0.76

thâu

-0.76

twimg

-0.74

 nahilalakip

-0.74

oa̍t

-0.74

IntoConstraints

-0.74

POSITIVE LOGITS

1.55

-...

0.59

-$

0.56

-£

0.54

-[

0.54

es

0.53

-]

0.50

-/

0.49

-<

0.48

Activations Density 0.010%

references to the concept of self

references to identity or self-concept

self- compounds

No Comments

No Known Activations

references to the concept of self

references to identity or self-concept

self- compounds

No Comments

No Known Activations