INDEX

Explanations

references to characters' speech patterns and accents

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

accent or speech pattern

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

 intended

-0.41

 weer

-0.36

 ממש

-0.36

nhof

-0.35

subject

-0.35

expected

-0.35

 continues

-0.35

textLabel

-0.35

 cukup

-0.35

POSITIVE LOGITS

ftagPool

0.72

 surla

0.63

 AssemblyCompany

0.57

 invokingState

0.54

 Sequelize

0.54

 OFDb

0.52

posedge

0.51

XMLSchema

0.51

twimg

0.50

 resourceCulture

0.50

Activations Density 0.056%

references to characters' speech patterns and accents

accent or speech pattern

No Comments

No Known Activations