INDEX

Explanations

coding terms related to file handling and access permissions

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

fileprivate var, #define, related to

np_acts-logits-general · gemini-2.5-flash-lite

New Auto-Interp

Configuration

google/gemma-scope-9b-pt-res/layer_24/width_131k/average_l0_97

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.24.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

 Normdatei

-1.08

GEBURTSDATUM

-1.02

 Administrativna

-1.02

ſſung

-1.01

 Biôgrafia

-0.98

<unused43>

-0.96

<unused41>

-0.95

<unused74>

-0.95

<pad>

-0.95

<unused20>

-0.94

POSITIVE LOGITS

filepath

0.98

↵↵

0.61

↵↵↵

0.52

 filepath

0.52

The

0.50

↵↵↵↵

0.49

↵↵↵↵↵↵

0.49

the

0.49

↵

0.48

<eos>

0.43

Activations Density 0.000%

coding terms related to file handling and access permissions

fileprivate var, #define, related to

No Comments

No Known Activations