INDEX

Explanations

instances of the word "end" and its variations in various contexts

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

termination sequences

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

 queſta

-0.98

 ſind

-0.94

$_(

-0.87

 ſch

-0.86

 ſche

-0.85

 ſta

-0.84

 ſtand

-0.82

transQ

-0.81

ſelf

-0.80

<unused14>

-0.79

POSITIVE LOGITS

↵

0.33

 mesma

0.28

ostante

0.26

 lecteur

0.26

0.25

↵↵

0.24

poptotic

0.24

meu

0.24

Empereur

0.24

Activations Density 0.002%

instances of the word "end" and its variations in various contexts

termination sequences

No Comments

No Known Activations