INDEX

Explanations

statements regarding predictions and interpretations of legal rulings

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

uncertainty or possibility

np_acts-logits-general · gemini-2.0-flash

likely, would, will, might, probability

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-16k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_16k/average_l0_91

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

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

 ſche

-0.52

 GenerationType

-0.50

 ſta

-0.48

 faſt

-0.45

getInteger

-0.44

 ſte

-0.44

 houſe

-0.43

 perſon

-0.42



-0.40

XmlAccessorType

-0.40

POSITIVE LOGITS

 likely

0.63

likely

0.63

 Likely

0.61

httphttps

0.59

Likely

0.58

probably

0.56

 Probably

0.55

Probably

0.54

 probably

0.54

 probablemente

0.52

Activations Density 0.428%

statements regarding predictions and interpretations of legal rulings

uncertainty or possibility

likely, would, will, might, probability

No Comments

No Known Activations