INDEX

Explanations

phrases or questions that inquire about the next steps or forthcoming actions

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

next event or step

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

 isolado

-0.48

middelen

-0.43

 Khusus

-0.42

 alone

-0.40

 tebal

-0.40

 strani

-0.39

 prijzen

-0.38

 marinho

-0.37

 pepper

-0.37

 considérable

-0.37

POSITIVE LOGITS

 next

1.05

Next

1.04

 Next

1.02

next

1.00

 NEXT

0.98

 getNext

0.93

NEXT

0.92

 prochaines

0.87

下一步

0.87

 nächste

0.86

Activations Density 0.024%

phrases or questions that inquire about the next steps or forthcoming actions

next event or step

No Comments

No Known Activations