INDEX

Explanations

terms related to legal contexts and processes

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 20-gemmascope-res-16k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_16k/average_l0_91

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

Packard

-0.64

 Jefus

-0.59

setcounter

-0.59

stalline

-0.58

Slf

-0.58

 faſt

-0.57

ImageContext

-0.57

actéristi

-0.57

 disulfide

-0.57

 Chrift

-0.56

POSITIVE LOGITS

 brief

0.56

 BRIEF

0.54

zerű

0.54

Brief

0.53

 Brief

0.50

brief

0.47

equivalent

0.46

Equivalent

0.46

rzez

0.45

 lediglich

0.43

Activations Density 0.242%

terms related to legal contexts and processes

No Comments

No Known Activations