INDEX

Explanations

the word "that" used in various contexts

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

describing something that

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

 Chwiliwch

-0.57

 onPostExecute

-0.47

 betweenstory

-0.46

 nakalista

-0.45

 Савезне

-0.43

 تكبرها

-0.43

تقاوى

-0.43

Kaieteur

-0.42

 heisst

-0.41

 mulighed

-0.40

POSITIVE LOGITS

 inevit

0.53

 inevitably

0.52

 invariably

0.49

 inevitable

0.49

 undoubtedly

0.46

 diver

0.45

 moods

0.45

 Peter

0.45

 Ellen

0.45

 always

0.45

Activations Density 0.023%

the word "that" used in various contexts

describing something that

No Comments

No Known Activations