INDEX

Explanations

references to specific entities and conditions related to investigations or research findings

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

conditional or hypothetical language using words like "that", "which", "can", "is", and "were".

oai_token-act-pair · claude-3-5-haiku-20241022 Triggered by @neilrathi

New Auto-Interp

Top Features by Cosine Similarity

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

Configuration

google/gemma-scope-9b-pt-res/layer_20/width_16k/average_l0_68

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

IBOutlet

-0.52

UpInside

-0.48

SBATCH

-0.40

+#+

-0.40

ientras

-0.40

LEGGI

-0.39

كويكب

-0.39

ReusableCell

-0.39

expliquer

-0.39

Портали

-0.38

POSITIVE LOGITS

Билгалдахарш

0.43

 Chwiliwch

0.42

wości

0.42

Sebastian

0.41

Rumah

0.40

 potenti

0.40

々が

0.40

 Sebastian

0.40

 affected

0.40

kaper

0.40

Activations Density 0.391%

references to specific entities and conditions related to investigations or research findings

conditional or hypothetical language using words like "that", "which", "can", "is", and "were".

No Comments

No Known Activations