INDEX

Explanations

conversational affirmations and responses

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

acknowledging understanding

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

 itſelf

-0.80

 ſche

-0.77

ſelf

-0.76

 themſelves

-0.73

 ſch

-0.72

 estekak

-0.71

 queſta

-0.71

ſelves

-0.69

 myſelf

-0.68

MLLoader

-0.66

POSITIVE LOGITS

 senhora

0.36

 dAtA

0.34

 Manns

0.33

 quindi

0.33

 forklar

0.31

0.29

 beliau

0.28

 Quindi

0.28

 explicou

0.28

Activations Density 0.015%

conversational affirmations and responses

acknowledging understanding

No Comments

No Known Activations