INDEX

Explanations

expressions of future intention or predictions

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

will be

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

 houſe

-0.63

 Houſe

-0.60

 AssemblyTitle

-0.59

 ſta

-0.58

 purpoſe

-0.56

⤹

-0.56

 pleaſure

-0.56

ſelf

-0.53

wiſe

-0.52

 Comprometido

-0.51

POSITIVE LOGITS

]]>

0.54

 segar

0.50

not

0.44

0.43

<th>

0.42

lin

0.41

 enlist

0.40

 become

0.40

be

0.39

Activations Density 0.024%

expressions of future intention or predictions

will be

No Comments

No Known Activations