INDEX

Explanations

phrases or constructions that indicate purpose or intention

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

purpose or context

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

-0.57

 trama

-0.49

-0.45

 consci

-0.44

 Nowak

-0.44

EndTime

-0.42

<eos>

-0.42

ed

-0.41

he

-0.41

POSITIVE LOGITS

for

1.02

FOR

1.00

for

0.98

 forState

0.98

 htmlFor

0.97

 endfor

0.97

For

0.97

For

0.93

 voor

0.90

สำหรับ

0.88

Activations Density 0.493%

phrases or constructions that indicate purpose or intention

purpose or context

No Comments

No Known Activations