INDEX

Explanations

references to old or older items and concepts

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

old or older

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

liceerd

-0.44

extAlignment

-0.41

TestRunner

-0.39

bollah

-0.39

 chấm

-0.39

 Biar

-0.36

secta

-0.35

isEqualTo

-0.35

IGraphics

-0.34

yticks

-0.34

POSITIVE LOGITS

old

0.72

 viejos

0.67

 vieux

0.67

 vieilles

0.65

 older

0.65

Older

0.62

古い

0.61

 vecchio

0.61

 oude

0.60

 古い

0.60

Activations Density 0.027%

references to old or older items and concepts

old or older

No Comments

No Known Activations