INDEX

Explanations

instances of duplication or redundancy in various contexts

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

counting, repetition, duplication

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

io

-0.38

aven

-0.34

inc

-0.30

rat

-0.30

iv

-0.30

el

-0.30

fit

-0.30

sprintf

-0.30

 appointed

-0.29

POSITIVE LOGITS

 kasarigan

0.94

WithIOException

0.85

 Roskov

0.83

aarrggbb

0.81

ValueStyle

0.81

ſelben

0.75

tagHelperRunner

0.75



0.74

enderror

0.74

 Infórmanos

0.74

Activations Density 0.060%

instances of duplication or redundancy in various contexts

counting, repetition, duplication

No Comments

No Known Activations