INDEX

Explanations

numerical values and their relationships in structured formats

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

user input or conversational turns

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

InputTagHelper

-0.45

RTGC

-0.43

 commission

-0.40

 negatively

-0.39

 ProtoMessage

-0.39

 nadzieję

-0.37

scriptId

-0.37

getColumnIndex

-0.37

weiterlesen

-0.37

 PLAST

-0.37

POSITIVE LOGITS

SizeF

0.52

 beginnetje

0.50

 חיצוניים

0.49

 consultato

0.49

 gawas

0.48

 Allociné

0.47

Réalisation

0.47

 الدولى

0.47

 defaultstate

0.45

 ब्रेकडाउन

0.45

Activations Density 0.011%

numerical values and their relationships in structured formats

user input or conversational turns

No Comments

No Known Activations