INDEX

Explanations

punctuation and specific keywords related to structure or events

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

Tokens followed by "and." or that contain "and"

np_acts-logits-general · gemini-2.0-flash

rand, rank, vand, war, warrand

np_acts-logits-general · gemini-2.5-flash-lite

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 22-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_22/width_16k/average_l0_72

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.22.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

 Riddle

-0.51

 Gleason

-0.49

 Twit

-0.49

 Pryor

-0.48

 Wiggins

-0.48

Coe

-0.48

 Higgins

-0.47

 Hoffman

-0.46

 Hickey

-0.46

copa

-0.44

POSITIVE LOGITS

 RAND

1.11

 Rand

1.09

 Kand

1.09

RAND

1.08

 rand

1.06

 Randall

1.06

rand

1.04

 kand

1.03

Rand

1.02

 Tand

1.02

Activations Density 0.655%

punctuation and specific keywords related to structure or events

Tokens followed by "and." or that contain "and"

rand, rank, vand, war, warrand

No Comments

No Known Activations