INDEX

Explanations

phrases related to understanding and knowledge

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-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

Cyfeiriadau

-0.38

 SIMBAD

-0.37

richTextPanel

-0.35

Superficie

-0.33

 Audiodateien

-0.33

CodedInputStream

-0.33

 håll

-0.32

 Drapeau

-0.32

TEMPO

-0.32

CrossRef

-0.32

POSITIVE LOGITS

 incompetence

0.72

 incompetent

0.60

 incompet

0.55

contentLoaded

0.52

 resourceCulture

0.48

 unskilled

0.48

compet

0.48

 inept

0.47

 ignorant

0.47

 компет

0.47

Activations Density 0.023%

phrases related to understanding and knowledge

No Comments

No Known Activations