INDEX

Explanations

markers and identifiers in structured data or code segments

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 38-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_38/width_16k/average_l0_128

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.38.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

stdc

-0.48

-0.43

vis

-0.42

 '\\;'

-0.40

 Vile

-0.39

 comple

-0.39

 kasarigan

-0.39

"))

-0.38

clientY

-0.38

Dil

-0.38

POSITIVE LOGITS

 nakalista

0.55

adaptiveStyles

0.53

 câte

0.49

 desmotivaciones

0.47

 henkil

0.47

tvguidetime

0.46

 oprot

0.46

 ostavi

0.45

harapkan

0.45

 automatiques

0.45

Activations Density 0.337%

markers and identifiers in structured data or code segments

No Comments

No Known Activations