INDEX

Explanations

terms related to disconnection or absence, particularly "out."

oai_token-act-pair · gpt-4o-mini

Triggered by @bot

out within words

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 9-gemmascope-res-131k

Configuration

google/gemma-scope-9b-it-res/layer_9/width_131k/average_l0_121

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.9.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

IFrame

Link

Negative Logits

 ainfi

-0.83

 dezelve

-0.76

 vectorielle

-0.74

 vectorielles

-0.67

 zoude

-0.65

 zijne

-0.63

vœ

-0.60

 quelcon

-0.60

Cordialement

-0.59

 médicale

-0.58

POSITIVE LOGITS

out

2.11

Out

1.64

OUT

1.56

Out

1.52

out

1.51

OUT

1.46

outs

1.45

 outs

1.28

EOUT

1.25

Outs

1.10

Activations Density 0.011%

terms related to disconnection or absence, particularly "out."

out within words

No Known Activations