INDEX

Explanations

occurrences of characters or events in a narrative context

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

pursuant to its agreement

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

 pleaſure

-0.69

ſelf

-0.68

 ſche

-0.64

 purpoſe

-0.62

 ſta

-0.60

ſelves

-0.59

 Houſe

-0.58

 faſt

-0.56

 ſtate

-0.55

 houſe

-0.54

POSITIVE LOGITS

BufferException

0.50

 iprot

0.46

 bounced

0.45

<()>

0.45

 sinned

0.43

 glanced

0.42

 finally

0.41

tagHelperRunner

0.41

 gasped

0.41

 blinked

0.40

Activations Density 0.018%

occurrences of characters or events in a narrative context

pursuant to its agreement

No Comments

No Known Activations