INDEX

Explanations

phrases indicating direction or goals

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

direction or progression

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

 enfans

-0.64

Edi

-0.47

 Majefty

-0.46

SEGUIR

-0.43

 luster

-0.42

deleteItem

-0.42

Susie

-0.41

 saja

-0.41

 Púb

-0.40

resultCode

-0.40

POSITIVE LOGITS

 towards

1.32

towards

1.26

 Towards

1.17

 toward

1.14

toward

1.08

Towards

1.07

 Toward

1.04

Toward

0.91

 hacia

0.82

 προς

0.81

Activations Density 0.028%

phrases indicating direction or goals

direction or progression

No Comments

No Known Activations